Guild Wars 2

Note: The below contains quick numbers which can easily be modified to be more realistic, I just threw this together quickly in Excel on a break. I also have no reference for the value I assigned to ‘r’.

The RNG code of the Mystic Forge can be changed so that extreme outliers are less frequent.
Let r = 0.001% (1 in 100 000), the current chance of getting a precursor through the MF
Let n = number of previous failed attempts at getting a precursor through the MF. It resets to zero upon a success.
Let f = 1.1, the increase in chance of getting a precursor due to previous failures
Let P(r,n,f) = Current probability of your account getting a precursor through the MF

Using P=r*(f^n), ~79% of accounts will have received at least one precursor by the time they had done 100 trials at the MF. This is in contrast to the 0.1% of accounts that would have received at least one precursor by the 100 trial mark.

Again, the values above are just a sample, and yes, extreme outliers will still exist as they always will, but they will exist less frequently.

Edit: Quick drop of code from Excel. Cells delineated with commas. Had to add a space after “PRODUCT” so the forum would show the full code.
Chance of REWARD, 0.00100%
Increase odds upon failure ,1.1

Trial Number, Chance of REWARD, Failure %, Running Chance
0, =$B$1*($B$2^A5), =1-B5, =1-C5
=A5+1, =$B$1*($B$2^A6), =1-B6, =1- PRODUCT ($C$5:C6)

Randulf.7614:

Brother Grimm.5176:

I’m going to get some popcorn…..

Might I add that a simple comment from John (or any Dev) that states emphatically that there is NO secret LUCK factor tied to any player / account / character anywhere in the code (or ever has been), might help to squash some of the crazyiness I envision showing up in this thread. Just a suggestion.

He quashed that emphatically in the other thread this span off from

Doesn’t mean that everyone who posts in this thread have read the other. Best to post it here too although most likely won’t read it once this thread has gotten long enough.

Having a method to reduce mainly the lower outliers would be beneficial, but you are adjusting a single layer of RNG.

Most other MMO’s have more simplified tables where if you want X item, you know that the following A, B, C monsters have them at a given chance; so you can farm for that key item. Guild Wars 2 has been built around a model that tries to discourage this with a giant loot table that goes something like this:

40% – Tier 1 (Some random Junk/Whites from some other table)
30% – Tier 2 (Some random Blues/Mats from some other table)
25% – Tier 3 (Some random Greens/Better Mats from some other table)
4% – Tier 4 (Some random Rares/Rare Mats from some other table)
0.9% – Tier 5 (Some random Exotic from some other table)
0.1% – Tier 6 (Some random Super Special item like a recipe or skin)

So say you finally hit a Tier 5 drop and get an exotic, but what kind of exotic? It could be anything from a piece of armor your class can’t even equip to, if you’re lucky, that weapon with the stat you wanted. I feel that’s a problem, because even if you guarantee a rare/exotic drop after a long dry spell, is that drop going to be useful? Is it worth anything more than just a salvage or market fodder? At the end of the day, you’re still left not feeling rewarded and maybe a little more aggravated because you still haven’t gotten an item you could have used.

Option 2 is hands down the superior option. Removing the chance for or mitigating low drops has too many long range repercussions. But for people like my who have uncanny bad luck in everything (not hosting a pity party here, this us just a genuine fact about me :P), there needs to be some kind of extra system to counteract that.

ShadowMageAlpha.7580:

I have been playing for Team Fortress 2 for quite some time, long before it was to become the war-like hat simulator that it is now.
When the hat system first came out, the method of getting drops for it was 100% random. It ended up that hats were INSANELY valuable because there was a significant portion of the people that would literally never get a hat drop. If I recall, I was one of those people and I played a LOT.
Valve ended up changing the system to a pseudo-RNG. After X amount of time, you were given a chance for a hat to drop, but you would ALWAYS be given a hat after Y amount of time. I’m not sure at the exact mechanics of the system, but it effectively moved all those hatless “brutally unlucky” people up into a more middle of the road “unlucky that they’re not more lucky” group. It almost universally increased the “class” of all the players, and the only group “suffered” were the lucky group of people that had a number of hats drop within a short amount of time. They only really “suffered” in a sense that their group became less exclusive and their hats became less valuable.
Long story short, Valve tried the pure RNG approach for a while, and something like 95% of the people or more were not happy with it. They changed to a pseudo-RNG system with a build up of luck as time passed, and those people became MUCH happier.

I don’t like this comparison between Team Fortress 2 and Guild Wars 2, something is amiss.

Ayrilana.1396:

Essence Snow.3194:

Ok then……why isn’t it true, give some examples….Saying no without adding anything doesn’t help anything.

Just look at whatever items in the game have low supply but high demand. Look at precursors. Whether something will have a low supply and high demand is not in the player’s control but with Anet. They have their own vision on what they would like for each item. Since precursor drop rates are very low, and they’re highly sought after, it’s pretty safe to assume that they intended for this imbalance between supply and demand.

you can simulate the exact same output from random through other means. in large numbers random is fairly predictable and known.
for example, it is highly likely their is a consistent number for the current random implementation that you can represent like this.
X amount of enemies killed per precursor drop.
which also translates into X amount of player hours per precursor drop.

once you realize this formula, it wouldnt be difficult to create a set of conditions that can simulate the exact same output.

random is a design choice, its not a necessity in terms of predetermining supply.

EnemyCrusher.7324:

You’re thinking about this way too narrowly. The non-randomness of John’s second suggestion would mean that after a certain number of unlucky drops, you’re guaranteed to get a lucky drop. This wouldn’t just apply to the mystic forge, but I’ll use that as my example since you did. If you throw 4 rare level 80 longbows into the mystic forge 20 times and only get rares, no exotics, on the 21st try you would be guaranteed an exotic. Whether that exotic is a precursor or not would still be up to luck. There would also be no actual “tickets”, that’s a figurative description for the guarantee of better luck on your next try.

My depiction also gets you a guaranteed item, but it has two significant benefits over a non-ticket guarantee system:

1. You can CHOOSE the item you want if you have failed enough times.
2. You can SEE your progress towards the item you want.

Both of those benefits will result in substantially happier players in general, which is why a ticket based system is superior to a behind the scenes improvement in your odds.

RNG is great. Everyone loves RNG. We need more and more of it.
When buying gems for 100€ there should be rng which decides if you get 0 or upto 100k gems!

OPtion 2 sounds great but the dependent factor to see if someone should get a loot ‘bump’ can’t be time or even time played. Perhaps it’s number of missed looting chances, further dependent on the difficulty of the mob? To implement something like this, a loot parameter would have to be tracked per character. You also have the complexity of a factor Anet already has to give bonus loot to people; Magic Find. That could also be another way to ‘bump’ a player that has been unlucky.

phys.7689:

Ayrilana.1396:

Essence Snow.3194:

Ok then……why isn’t it true, give some examples….Saying no without adding anything doesn’t help anything.

Just look at whatever items in the game have low supply but high demand. Look at precursors. Whether something will have a low supply and high demand is not in the player’s control but with Anet. They have their own vision on what they would like for each item. Since precursor drop rates are very low, and they’re highly sought after, it’s pretty safe to assume that they intended for this imbalance between supply and demand.

you can simulate the exact same output from random through other means. in large numbers random is fairly predictable and known.
for example, it is highly likely their is a consistent number for the current random implementation that you can represent like this.
X amount of enemies killed per precursor drop.
which also translates into X amount of player hours per precursor drop.

once you realize this formula, it wouldnt be difficult to create a set of conditions that can simulate the exact same output.

random is a design choice, its not a necessity in terms of predetermining supply.

I think you quoted the wrong post of mine as your post doesn’t have anything to do with what you quoted of mine. I’ll go ahead and answer you anyway.

You’re assuming the determination of the drop rate is based on player action rather than a manually added percentage. I may have read your post wrong. If i did, could you clarify.

John Smith.4610:

This is a spinoff of the economy thread to talk about RNG tactics in games in a general form.
Here’s the premise. RNG is evenly distributed on aggregate. On an individual level this means that while almost everyone falls into a reasonable range in the middle, there are outliers on each side of the distribution that are either highly rewarded or not rewarded at all. These individuals become sample cases and spotlights for experiences that maybe shouldn’t exist.

We do need to be very careful about ideas that flatten the experience entirely as that quickly becomes not fun at all.

There are two concepts that have been discussed in the other thread that I’ll briefly summarize.

1. Use a specifically non-random NG. The NRNG functions similarly to a RNG, but has characteristics that either squish the distribution so that outliers exist much less or specifically manipulate a player’s experience for loot in a more complicated way that makes it feel rewarding.

2. Implement measures that counteract low-end outlier behavior inside of game design. This would be a system that is something like: If player hasn’t received a rare drop in X time send them Y tickets for random drops.

Obviously these are hyper-simplified descriptions, but I don’t want this to get too long.

Thanks for the information. I guess it takes a lot of courage to admit that there is some kind of luck account.

If you are talking about normal distribution, lower the standard deviation = i choose the first option. The second option means more difficulties as it means a more complex formula although the first option could disturb the market price equilibriums.

Obtena.7952:

OPtion 2 sounds great but the dependent factor to see if someone should get a loot ‘bump’ can’t be time or even time played. Perhaps it’s number of missed looting chances, further dependent on the difficulty of the mob? To implement something like this, a loot parameter would have to be tracked per character. You also have the complexity of a factor Anet already has to give bonus loot to people; Magic Find. That could also be another way to ‘bump’ a player that has been unlucky.

Excellent post.

I’m all for 2.

Based on the many games I’ve played and all the feedback I’ve read from countless players, what will definitely work is 2:

• If you are lucky, nothing happens, good for you, go on.
• If you are unlucky for many drops generated for your account, the “safety” kicks in and your chances of getting something rare increase each time a drop is generated for you with “unlucky” stuff, until you finally getting the stuff you are looking for is inevitable.
• Regardless of you being lucky ‘naturally’ or through he safety system, when the stuff finally drops, the safety resets.

Every time such a ‘cushion’ is added, less people feel like they are never getting anything or that the only way to get anything is farming then buying the stuff from the few that do get it.

VitaminK.2517:

Thanks for the information. I guess it takes a lot of courage to admit that there is some kind of luck account.

GRIMM! Where is is Brother Grimm!!?

What about a Pseudorandom number generator (PRNG)?

This keeps the RNG the same as it is but helps the people on the lower end as every time they fail to get rewarded with awesome loot it will keep increasing their chance until they eventually get the awesome loot.

MithranArkanere.8957:

I’m all for 2.

Based on the many games I’ve played and all the feedback I’ve read from countless players, what will definitely work is 2:

• If you are lucky, nothing happens, good for you, go on.
• If you are unlucky for many drops generated for your account, the “safety” kicks in and your chances of getting something rare increase each time a drop is generated for you with “unlucky” stuff, until you finally getting the stuff you are looking for is inevitable.
• Regardless of you being lucky ‘naturally’ or through he safety system, when the stuff finally drops, the safety resets.

Every time such a ‘cushion’ is added, less people feel like they are never getting anything or that the only way to get anything is farming then buying the stuff from the few that do get it.

I’m for this method.

Edit: only problem is, how would this effect such things as the world boss chests that garuntee a rare or better?

Ayrilana.1396:

phys.7689:

Ayrilana.1396:

Essence Snow.3194:

Ok then……why isn’t it true, give some examples….Saying no without adding anything doesn’t help anything.

Just look at whatever items in the game have low supply but high demand. Look at precursors. Whether something will have a low supply and high demand is not in the player’s control but with Anet. They have their own vision on what they would like for each item. Since precursor drop rates are very low, and they’re highly sought after, it’s pretty safe to assume that they intended for this imbalance between supply and demand.

you can simulate the exact same output from random through other means. in large numbers random is fairly predictable and known.
for example, it is highly likely their is a consistent number for the current random implementation that you can represent like this.
X amount of enemies killed per precursor drop.
which also translates into X amount of player hours per precursor drop.

once you realize this formula, it wouldnt be difficult to create a set of conditions that can simulate the exact same output.

random is a design choice, its not a necessity in terms of predetermining supply.

I think you quoted the wrong post of mine as your post doesn’t have anything to do with what you quoted of mine. I’ll go ahead and answer you anyway.

You’re assuming the determination of the drop rate is based on player action rather than a manually added percentage. I may have read your post wrong. If i did, could you clarify.

Point is, anet can generate the same supply without random, if they choose to do so.

The advantages of random are not really about controlling supply, its mostly the benefits of unpredictability

phys.7689:

Brother Grimm.5176:

…
I’m pretty sure that nobody think there is special code about that.

But we’re a lot to rather think that RNG is bugged in some way, and that’s why, in addition to a “gauss distribution” as described in the previous post, some players are always lucky and some are always unlucky. As the RNG mechanism is all in all quite complex (modified by magic find ; depends on damage done by the player / by his group compared to damage done by other players ; special items with a different table loot ; …), there could be several bugs introduced in the system. Remember that the % of magic find displayed in the hero panel is wrong since months (but the real number is supposed to be correctly applyed), so it wouldn’t be surprising that there are some bugs in the loot mechanism itself.

There are CERTAINLY very vocal players that are convinced of the “secret LUCK stat”.

its not really relevant weather there is a secret luck stat or not. If a random works properly, there will be people who will, will seem lucky, or those who seem unlucky, when compared to the data. This is the normal expectation of the bell curve.

So whether it is a bug or not doesnt matter. The point is these people are bound to exist.

My take on RNG, isn’t that it exist or it’s broken, it’s just that it’s practically everywhere. Neither of those solutions would work anyway, since the game would have to predict what you’re after in the first place, then reward you for failure of what exactly? Do we put a merchant in the game that carries everything a player might want in exchange for a loser token?

Here are my observations of the problems…

First and foremost, stop with purchases in gem shop that contain RNG. I don’t want to play grab bag to get a skin or item i want. Sell me the dye i want, not a random chance at it. Same things with the the tickets. I don’t even mind the chests all that much, just keep it to items you’d normally could or can currently buy in the store anyway.

Second, stop with the clusters and amberite type nonsense. Clusters at least we can buy, but still, the drop rate is horrible. I’m definitely not saying you guys should just hand everything players want over, but at least give us ways to actually earn them.

Third, the forge, it’s actually perfectly fine aside from precursors. I won’t go into the whole crafting bit.

Loot drops and chests are also fine, although i really wish there was more diversity.

That’s my 2 cents.

I’d like to start by saying that RNG itself is not the issue here. The main issue (as far as I’ve been able to suss out) is that this game has very, very few ways to decide what type of RNG you’re up against.

Here’s the comparison that makes the most sense to me, since it’s drawn from years of experience: in WoW (yes, that game, but where it’s from is not the point) there are various levels of RNG. Most of the time, if you want a specific gear item you can look it up and see that it only drops from certain bosses, often in the 10-25% range. These are what you could consider typical items, not meant to be exclusive. Other items are meant to be difficult to obtain, and so have drops in the 0.01-2% range, but are still dropped by specific mobs. Then, you have what are known as “world drops”. These are items that have often a 0.05-2% drop range, AND have a very wide pool of mobs it can drop from, often across multiple zones. In other words, for “typical” and most “exclusive” items, you can mitigate the RNG by knowing specifically what to kill, thereby limiting your pool of possible drops. For “world drops”, you essentially cannot mitigate by limiting the item pool.

The underlying problem in GW2 is that almost all desirable specific drops, including exotics with unique skins, are “world drops” or very close to it. Only in a few specific instances can you limit your item pool, as with the Sunless weapon chest. Otherwise, there is no way to specifically “go after” or hunt for what you want – at level 80, you could have a precursor drop from the gate guardian of Arah… or any random Risen mob… or a level 5 fish in Queensdale… or from a champ box… or in a dungeon… or from a JP or other chest. You get the idea. Now, I’m not suggesting that each precursor should have a specific list of mobs it drops from since maybe that’s asking too much, but I definitely would have no problem with other unique skins or more “typical” gear and vanity items having more specific methods of acquisition than a completely random world drop. The chance to drop doesn’t even have to be radically better, it just needs to give you a way to feel like you’re on the right track so to speak. It would also be nice if some more common or less prestigious skins could have more like the 10-25% drop rate, say skins from dungeon bosses.

contrary to munkiman I love the ambrite drops compared to the mystic forge. You get something while actually doing something in the world. The drops aren’t too bad either, got my first one after a few chests. Even if the droprate would be bad, you’d still have to play the game to get the keys in order to unlock the chests. This is imho good game design.

The mystic forge on the other hand is just a playground for (mostly) players who already play little open world vs. trading post. If you have tons of money you can buy tons of stuff which you throw in the mystic forge. I rarely get enough rare greatswords in a short time in order to try my luck on dusk/dawn. It takes time to get 4 of the same rare weapons in the world, in the trading post… not so much.

@John Smith:

RNG was fairly bad in other games, but it’s far worse in GW2 because while in other games the loot roll is for the group, in GW2 there is an individual roll for each person.

This eliminates the concept of luck sharing based upon preference.

When a end-tier light armor chestpiece dropped for the group you wouldn’t have to compete with heavy/medium users, or with those who already have that chestpiece. In GW2 the game rolls 5 times and if you happen to get the item it takes no consideration of if you actually want it or not.

For tradable items this can be argued to be ‘fair’, as you can just sell your drops over time and purchase the wanted items- even though this greatly diminishes the ability of players to directly link a sense of achievement with their items.

But for untradable items, personal RNG is a monster that sucks all the fun from any activity it touches.

“kitten another <the item you desperately want>, guess I’ll need to vendor it for 5 silver.” Is a sentence that you should never, ever- ever, allow your players to ever hear. The fact such a sentence can exist is a kitten ing indicator that the system needs to change.

phys.7689:

Ayrilana.1396:

phys.7689:

Ayrilana.1396:

Essence Snow.3194:

Ok then……why isn’t it true, give some examples….Saying no without adding anything doesn’t help anything.

Just look at whatever items in the game have low supply but high demand. Look at precursors. Whether something will have a low supply and high demand is not in the player’s control but with Anet. They have their own vision on what they would like for each item. Since precursor drop rates are very low, and they’re highly sought after, it’s pretty safe to assume that they intended for this imbalance between supply and demand.

you can simulate the exact same output from random through other means. in large numbers random is fairly predictable and known.
for example, it is highly likely their is a consistent number for the current random implementation that you can represent like this.
X amount of enemies killed per precursor drop.
which also translates into X amount of player hours per precursor drop.

once you realize this formula, it wouldnt be difficult to create a set of conditions that can simulate the exact same output.

random is a design choice, its not a necessity in terms of predetermining supply.

I think you quoted the wrong post of mine as your post doesn’t have anything to do with what you quoted of mine. I’ll go ahead and answer you anyway.

You’re assuming the determination of the drop rate is based on player action rather than a manually added percentage. I may have read your post wrong. If i did, could you clarify.

Point is, anet can generate the same supply without random, if they choose to do so.

The advantages of random are not really about controlling supply, its mostly the benefits of unpredictability

I brought that up in a previous post. Supply of an item will increase greater under a system with a guaranteed than one on RNG. When players have a guaranteed way, they will flock to that way. Those with a lot of free time will farm that way nonstop flooding the market. With an RNG system, this would not happen.

John Smith.4610:

munkiman.3068:

phys.7689:

Brother Grimm.5176:

…
I’m pretty sure that nobody think there is special code about that.

But we’re a lot to rather think that RNG is bugged in some way, and that’s why, in addition to a “gauss distribution” as described in the previous post, some players are always lucky and some are always unlucky. As the RNG mechanism is all in all quite complex (modified by magic find ; depends on damage done by the player / by his group compared to damage done by other players ; special items with a different table loot ; …), there could be several bugs introduced in the system. Remember that the % of magic find displayed in the hero panel is wrong since months (but the real number is supposed to be correctly applyed), so it wouldn’t be surprising that there are some bugs in the loot mechanism itself.

There are CERTAINLY very vocal players that are convinced of the “secret LUCK stat”.

its not really relevant weather there is a secret luck stat or not. If a random works properly, there will be people who will, will seem lucky, or those who seem unlucky, when compared to the data. This is the normal expectation of the bell curve.

So whether it is a bug or not doesnt matter. The point is these people are bound to exist.

My take on RNG, isn’t that it exist or it’s broken, it’s just that it’s practically everywhere. Neither of those solutions would work anyway, since the game would have to predict what you’re after in the first place, then reward you for failure of what exactly? Do we put a merchant in the game that carries everything a player might want in exchange for a loser token?

This was the first problem I thought of as well. It’s simple to say failure and success, but that’s not an easily definable concept in real life.

The loser token is a simple solution,

You can create a hunter system, whereby drop.rate of certain items of which you select one, have the behavior of increasing chance/streak breaking

You can make it so the mystic forge has methods of altering drops (forge can change weapon types, stat distributions for example)

There are a number of ways to skin a cat. However I think the best way is to design it into the game design of rare items that their is a more predictable alternate method of obtaining them

Mister Smith, what is the purpose of this thread? Is it to educate us about how RNG work on online games? Is it about making suggestions to improve the reward systems on the game? both? none?
I can see that everyone is giving their opinion on what is better and what doesn’t work, but to me is not clear the goal of this thread, and I would like to know.

Said that, I will just give a few comments on what is being discussed:

- Fringe cases like the super lucky or the ultra bad luck are, in my opinion, part of the randomness that RNG systems generate, and without them it wouldn’t really be random. This is like the experiment of throwing a coin multiple times vs imagining you throw a coin multiple times. People tend to think (on the virtual case) that this kind of extreme situations “shouldn’t” happen, which is wrong, specially with independent events like throwing a coin, or killing a monster. If the goal is to have a RNG system, then this is not a problem.

- I don’t like half of the reward system of the game. I’m talking about the RNG part of it.
This is weird for me, because I tend to enjoy games where loot is based on RNG, but again, I think GW2 suffer from a similar problem the game Diablo 3 suffered during its initial months.

• We get too much loot
• Most of the time, loot is not interesting or compelling. Salvage fodder.
• This make players tend to prefer content that gives stable income over gambling by doing content that takes more time or effort (basically because by doing that kind of content, you don’t even get better rewards).
• The value of loot is too “gold-centric”, it should be, but at the same time, since nobody expects to get useful gear or items from loot, everyone sees the items dropped as their gold value.
• Initially, the idea of being able to do anything you want and be rewarded for that, is very attractive. But in reality, this means that nothing gives valuable rewards, and there is no reason to do any content over other. This make players think too much about efficiency, instead of challenge (neither of both is bad, but it should be balanced between the two concepts).

This is for the RNG part of the reward system. I like the other systems, and one in particular, helps a lot while dealing with the RNG part, I’m talking about sPvP reward tracks, this system seems to me that is the best approach given the nature of the game. (The more we progress, we move between RNG boxes and certain rewards, this are themed, so it has meaning choosing one reward track over another.)

I wrote this post with a robotic tone, sorry for that.

phys.7689:

John Smith.4610:

munkiman.3068:

phys.7689:

Brother Grimm.5176:

…
I’m pretty sure that nobody think there is special code about that.

But we’re a lot to rather think that RNG is bugged in some way, and that’s why, in addition to a “gauss distribution” as described in the previous post, some players are always lucky and some are always unlucky. As the RNG mechanism is all in all quite complex (modified by magic find ; depends on damage done by the player / by his group compared to damage done by other players ; special items with a different table loot ; …), there could be several bugs introduced in the system. Remember that the % of magic find displayed in the hero panel is wrong since months (but the real number is supposed to be correctly applyed), so it wouldn’t be surprising that there are some bugs in the loot mechanism itself.

There are CERTAINLY very vocal players that are convinced of the “secret LUCK stat”.

its not really relevant weather there is a secret luck stat or not. If a random works properly, there will be people who will, will seem lucky, or those who seem unlucky, when compared to the data. This is the normal expectation of the bell curve.

So whether it is a bug or not doesnt matter. The point is these people are bound to exist.

My take on RNG, isn’t that it exist or it’s broken, it’s just that it’s practically everywhere. Neither of those solutions would work anyway, since the game would have to predict what you’re after in the first place, then reward you for failure of what exactly? Do we put a merchant in the game that carries everything a player might want in exchange for a loser token?

This was the first problem I thought of as well. It’s simple to say failure and success, but that’s not an easily definable concept in real life.

The loser token is a simple solution,

You can create a hunter system, whereby drop.rate of certain items of which you select one, have the behavior of increasing chance/streak breaking

You can make it so the mystic forge has methods of altering drops (forge can change weapon types, stat distributions for example)

There are a number of ways to skin a cat. However I think the best way is to design it into the game design of rare items that their is a more predictable alternate method of obtaining them

Like the PvP reward tracks in a sense?

In regards to 2.5, we already have that system. It’s the gold economy. Money is our “secondary reward” for not getting an interesting drop, and it is the resource we use to buy items we DO want.

However, that is crippled a bit in the current state of the game, mostly because money as a reward no longer has the range of uses it should. Other “secondary currencies” (and even tertiary currencies) have been implemented in newer areas, making money only really useful when 1) buying small goods, or 2) saving for an item that’s on the trade post.

I see the main problem here is that there are so many items that just can’t be bought with money, for various reasons, and the sheer shortage of “rewarding” items that can be bought with money. Yes, there are rewarding items to be bought on the trade post, but the items that many consider truly rewarding are at extremely high prices, due to supply and demand. If there were ways to infinitely generate these “rewarding items” (i.e., a vendor shop), then we wouldn’t have that issue.

However, we’ve been moving away from gold as a usable currency since the game started. First we had karma and dungeon tokens, then we got the commendation badges for PvP, then fractal relics, then SKILL POINTS as a currency, and now we’re up to geodes with Dry Top.

Why do we NEED these tertiary currencies? Why couldn’t the Dry Top goods be sold for gold? What’s the purpose of adding a tertiary system besides time-gating you into spending “x” amount of time in an area before you’re ALLOWED to by the thing you’re there for (and then slapping RNG on top of that with drop ratios on the currency itself)?

I can understand putting some of the rarer items behind a gate (high end recipes and the like), but I don’t see any reason for every item to be gated in such a way…

My biggest issue with the RNG of GW2 is that it’s almost win the jackpot or nothing (just to illustrate).

To me it’s a much better concept to have those Black lion ticket scraps so you work together a full ticket and feel the progress instead of having an ultrarare full ticket.

In the way the PvP-reward tracks work great. You actually work towards that dungeon armor chest instead of just having a very slim chance for it to drop. (Maybe compare to a ascended chest in Fractals – I Think the PvP version is better.)

To me it gets very boring with all the ascended rings I get from Fractals that I have no use for since i have done Fractals so much. I can’t do anything useful with them – not even salvage for a slim chance of ascended mats – and the money I get from selling them to vendor is close to a joke. It would actually be better to get a rare item instead. On the other hand the Fractal relics are great (even if they too overflood after a while).

Precursors are like the tip of the iceberg but illustrates the RNG problem very well. I have 3 750 hours of gameplay now and not gotten any precursor as a drop (but well from the Forge and the infamous early Karkaevent where it felt you rewarded all that didn’t get dc with a precursor). Sure, maybe I shouldn’t have gotten a precursor as drop in 3 750 hours anyway but it would have felt nice to work towards it instead of just having a 1 or 0, success or failure.

Luckystreaks as some have talked about can be a nice way I guess. But what is really a success from the Forge or from a Champion bag? After all, there are more interesting drops in this game than just precursors, at least to me, so who decides what Counts as a success and as a failure? Sure, you could have something like no-rare gives a temp bonus to MF – but then we will see people starting to playing that system too and I don’t think that would be good either. (Edit, yes, I know MF doesn’t affect champions chest or the Forge in the current system – just to illustrate…)

Anyway, i would love seeing solutions like better chance for a Sam / Foulbear kill without one, or maybe as a reward after killing 10k ogres. Or that Pendant of arah buyable for 5 k Arah token. Super-rare jackpots doesn’t trigger me as much as something I work towards where I feel a sence of progression.

Sorry for the long post.

John Smith.4610:

This is a spinoff of the economy thread to talk about RNG tactics in games in a general form.
Here’s the premise. RNG is evenly distributed on aggregate. On an individual level this means that while almost everyone falls into a reasonable range in the middle, there are outliers on each side of the distribution that are either highly rewarded or not rewarded at all. These individuals become sample cases and spotlights for experiences that maybe shouldn’t exist.

We do need to be very careful about ideas that flatten the experience entirely as that quickly becomes not fun at all.

There are two concepts that have been discussed in the other thread that I’ll briefly summarize.

1. Use a specifically non-random NG. The NRNG functions similarly to a RNG, but has characteristics that either squish the distribution so that outliers exist much less or specifically manipulate a player’s experience for loot in a more complicated way that makes it feel rewarding.

2. Implement measures that counteract low-end outlier behavior inside of game design. This would be a system that is something like: If player hasn’t received a rare drop in X time send them Y tickets for random drops.

2.5: “Add secondary reward mechanisms (ie. token based system) alongside the primary RNG system; allow progress to be made even when you don’t get the result you want.”

Obviously these are hyper-simplified descriptions, but I don’t want this to get too long.

edit: added 2.5

My vote would be for 2.5. We already have that with fractals and ascended rings (try your luck 10 times, if it didn’t pan out, well you can get it now anyway).

Thanks for making this thread, John. Looks to be an interesting discussion.

2.5 is something I’ve seen in other games and it certainly has its share of fans. I like the concrete aspect of it myself; no matter how bad your luck is, you’re eventually going to get what you came for in some capacity.

However, I’ve also noticed drawbacks to a token-based system:

1) Some people wind up looking at the system as inherently token-based and ignore the RNG component entirely. This gives them reason to complain about how many tokens are needed or to feel overwhelmed because they can calculate the time needed concretely to acquire desired items.

2) Acquiring every desired item connected to the tokens requires a finite amount of time (or “kills,” if we are talking about something like a boss that drops tokens). This means that (in theory) players are consistently going to hit the “I have nothing left to do” point and quit doing the activity. When RNG is present in peoples’ minds, it can keep them on the treadmill because “big reward could be waiting around the corner.” (Not unlike gambling.)

In summary: If the token system takes precedence in peoples’ minds, then they are more likely to ignore the RNG part entirely and thus, are more likely to quit a particular activity (independent of individual player luck) after a (somewhat predictable) finite number of times completing the activity.

Of course, that is looking at the token system in a vacuum. Pure RNG has its own set of disadvantages, such as people getting burned out on the lack of any discernible rewards and quitting an activity because they feel it is not worth the effort.

I think the main difference in disadvantage, in that regard, is that with a token system, you can largely predict when people are going to start losing interest and that it’s going to happen consistently at X point. Whereas with RNG, people will more often range from losing interest at any moment to doing an activity an infinite (that is, not measurably finite) number of times.

I’m not sure which system is more advantageous out of the two, or if something else is needed (such as one of the other systems you named). I will be watching the discussion and add more if I have other thoughts.

Ayrilana.1396:

phys.7689:

Ayrilana.1396:

phys.7689:

Ayrilana.1396:

Essence Snow.3194:

Ok then……why isn’t it true, give some examples….Saying no without adding anything doesn’t help anything.

Just look at whatever items in the game have low supply but high demand. Look at precursors. Whether something will have a low supply and high demand is not in the player’s control but with Anet. They have their own vision on what they would like for each item. Since precursor drop rates are very low, and they’re highly sought after, it’s pretty safe to assume that they intended for this imbalance between supply and demand.

you can simulate the exact same output from random through other means. in large numbers random is fairly predictable and known.
for example, it is highly likely their is a consistent number for the current random implementation that you can represent like this.
X amount of enemies killed per precursor drop.
which also translates into X amount of player hours per precursor drop.

once you realize this formula, it wouldnt be difficult to create a set of conditions that can simulate the exact same output.

random is a design choice, its not a necessity in terms of predetermining supply.

I think you quoted the wrong post of mine as your post doesn’t have anything to do with what you quoted of mine. I’ll go ahead and answer you anyway.

You’re assuming the determination of the drop rate is based on player action rather than a manually added percentage. I may have read your post wrong. If i did, could you clarify.

Point is, anet can generate the same supply without random, if they choose to do so.

The advantages of random are not really about controlling supply, its mostly the benefits of unpredictability

I brought that up in a previous post. Supply of an item will increase greater under a system with a guaranteed than one on RNG. When players have a guaranteed way, they will flock to that way. Those with a lot of free time will farm that way nonstop flooding the market. With an RNG system, this would not happen.

You are oversimplifying supply. You can design.systems that produce the same behavior.
If precursor from mob drop is .00001 chance you can make precursors drop from killing 100,000 monsters.
If the average player hours in combat is 5000, you can make it take 10,000 runs of a half hour activity

If 200 precursors are generated per day, you can create content that awards to the top 200 scorers (limit pet account per month)

Point is for any randomized supply there exists a direct predictable method to provide the same supply with the same input

John Smith.4610:

munkiman.3068:

phys.7689:

Brother Grimm.5176:

…
I’m pretty sure that nobody think there is special code about that.

But we’re a lot to rather think that RNG is bugged in some way, and that’s why, in addition to a “gauss distribution” as described in the previous post, some players are always lucky and some are always unlucky. As the RNG mechanism is all in all quite complex (modified by magic find ; depends on damage done by the player / by his group compared to damage done by other players ; special items with a different table loot ; …), there could be several bugs introduced in the system. Remember that the % of magic find displayed in the hero panel is wrong since months (but the real number is supposed to be correctly applyed), so it wouldn’t be surprising that there are some bugs in the loot mechanism itself.

There are CERTAINLY very vocal players that are convinced of the “secret LUCK stat”.

its not really relevant weather there is a secret luck stat or not. If a random works properly, there will be people who will, will seem lucky, or those who seem unlucky, when compared to the data. This is the normal expectation of the bell curve.

So whether it is a bug or not doesnt matter. The point is these people are bound to exist.

My take on RNG, isn’t that it exist or it’s broken, it’s just that it’s practically everywhere. Neither of those solutions would work anyway, since the game would have to predict what you’re after in the first place, then reward you for failure of what exactly? Do we put a merchant in the game that carries everything a player might want in exchange for a loser token?

This was the first problem I thought of as well. It’s simple to say failure and success, but that’s not an easily definable concept in real life.

If my understanding is correct, as far as the loot tables are concerned precursors are just named exotics. Thus, as far as the RNG is concerned you “win” if you get a named exotic whether it’s a precursor or not.

The reason why getting precursors are the most frequent example of RNG hate used is because players don’t equate other named exotics with precursors. Just as any precursor as a reward is far different than the precursor you want, or at least a valuable one. For example, I expect there would be quite a bit of rage if it were announced that everyone who hadn’t received a precursor in 3000 hours of playing would be sent a Rage via the mail to offset their bad RNG luck.

I like suggestion 2.5

A few solutions I think might work:

• Give players more control over loot tables. Let us narrow a bit the kind of items we can get. There is an example ingame right now: Rare Crafting Materials. I know, for example, that griffons of level 75+ will have a chance of dropping blood type and claw type materials, while barracudas of level 80 will have a chance of dropping scale type and fang type (if I’m not mistaken…) mats. This allow me to hunt for certain items, while at the same time being subject of RNG. Sadly this method is inefficient, because is always better to farm gold directly, and buy the mats.

• If meaningful drops are added in the future, think about adding systems similar to sPvP reward tracks to PvE. In fact, the same system should be working now on dungeons, instead of the token system. Is far more bearable to have loot boxes (RNG) as a bonus when you keep getting fixed rewards in between. (this is more of a perception thing, I know we get boxes and fixed rewards anyway from dungeons)

• Since rarity of gear is irrelevant in the drop system (once you are 80, you won’t be wearing greens, blues or whites, only exotics, and ascended are just too rare, like winning the lottery) Stat combinations should be the meaningful way of having control over our drops. Gear with certain stat combinations should be tied to certain themes on PvE (same for sPvP, WvW being the exception). This should lead to the addition of more creative stat combinations on exotics (2 stats only, 4 stats, new stats?, rune type of effects tied to the gear, etc)

• Super rare drops should be sellable, not-that-rare ones we can start thinking about making them account-bound. This will make doing certain content necessary (like you’ve done with Dry Top). My problem is, if gold is all we need to acquire things in the game, then efficiency will always be the best way of playing it, and ironically, or maybe good for you, the most efficient way is not playing the game, but playing with the credit card. I’m not against the gemstore or the idea of gems at all by the way, I think this is a problem with how gold-centric the rewards are, and how RNG works. Playing the game should always be more profitable than not playing it. (At least that’s what we expect from a non Pay to Win game)

1 and 2 would be very hard to implement I presume. Probably it will take ages for either of the 2 options to get in the game, so I vote no.
On the other hand the easiest and simple solution is 2.5 A token based system will allow everyone to benefit from it troughout normal gameplay. And this will encourage players to play even more. Cause they know the thing they are after is there, regardless of what time it would take to acquire. For instance Aion has such system. Hundreds of armor and weapon pieces, tools and whatnot available trough tokens AND normal drops of course. And it’s working great. Lots of encouragement to play even more.

Leave RNG as is maybe. I see a lot of people like to gamble. I myself am not a gambler, but that doesn’t keep me from buying gems for cool additions. The MF tho, now that’s a real pain. I was lucky to pull out 20+ precursors, but on the other hand I lost so much as well. I believe if you are persistant you will get lucky eventually. But there are days…

phys.7689:

Ayrilana.1396:

phys.7689:

Ayrilana.1396:

phys.7689:

Ayrilana.1396:

Essence Snow.3194:

Ok then……why isn’t it true, give some examples….Saying no without adding anything doesn’t help anything.

Just look at whatever items in the game have low supply but high demand. Look at precursors. Whether something will have a low supply and high demand is not in the player’s control but with Anet. They have their own vision on what they would like for each item. Since precursor drop rates are very low, and they’re highly sought after, it’s pretty safe to assume that they intended for this imbalance between supply and demand.

you can simulate the exact same output from random through other means. in large numbers random is fairly predictable and known.
for example, it is highly likely their is a consistent number for the current random implementation that you can represent like this.
X amount of enemies killed per precursor drop.
which also translates into X amount of player hours per precursor drop.

once you realize this formula, it wouldnt be difficult to create a set of conditions that can simulate the exact same output.

random is a design choice, its not a necessity in terms of predetermining supply.

I think you quoted the wrong post of mine as your post doesn’t have anything to do with what you quoted of mine. I’ll go ahead and answer you anyway.

You’re assuming the determination of the drop rate is based on player action rather than a manually added percentage. I may have read your post wrong. If i did, could you clarify.

Point is, anet can generate the same supply without random, if they choose to do so.

The advantages of random are not really about controlling supply, its mostly the benefits of unpredictability

I brought that up in a previous post. Supply of an item will increase greater under a system with a guaranteed than one on RNG. When players have a guaranteed way, they will flock to that way. Those with a lot of free time will farm that way nonstop flooding the market. With an RNG system, this would not happen.

You are oversimplifying supply. You can design.systems that produce the same behavior.
If precursor from mob drop is .00001 chance you can make precursors drop from killing 100,000 monsters.
If the average player hours in combat is 5000, you can make it take 10,000 runs of a half hour activity

If 200 precursors are generated per day, you can create content that awards to the top 200 scorers (limit pet account per month)

Point is for any randomized supply there exists a direct predictable method to provide the same supply with the same input

First off, something having a 0.0001% does not mean you’ll get it after 100,000 monster kills. If you were to set it exactly at 100,000 monster kills, you’d be cutting off the entire right half of the distribution curve and shifting those that would fall under it to the middle. This produces more supply as now those people acquire the items sooner.

It’s for that very reason that you’d be eliminating the probability of getting an item above the average which is why supply would increase more.

I’ve been working in Dry Top to get all the ambrite weapons, and yes, it took a long time, but I noticed something: it felt good to have a goal that I could work towards, see improvement, and know when I’d be done. There’s an element of rng in the insect drops, but getting the recipes wasn’t. I had to grind a lot to get geodes, but I knew I was making progress. With a precursor, I don’t. It’s just waiting for a drop, or making enough gold to buy it. I don’t think that’s satisfying. I think we need a better way then RNG. I think building up ‘luck’, for lack of a better term, is a good idea. I’ve been doing teq for the past 45+ days, and I haven’t gotten a sunless weapon. A guildie did it twice, and got one. I know people have been doing it longer and not gotten a sunless weapon. How is this helpful to people? I don’t think everyone should get a sunless weapon after the second time, but I think there has to be a better way then doing the event every day, hoping that you’ll get a rare drop, only to be disappointed when you don’t. This is demoralizing, and that does NOT help the game. It only hurts it.

I think the token solution works perfectly in the case of rare / exotic weapons into the Mystic Forge. I don’t think it’s necessarily appropriate everywhere.

I think things like Geodes work well for Dry Top currencies. The problem is multiple there, though. The Geode being an item is the first problem, since then it takes inventory space. What should have been done is to have it be an actual currency, and then remove the “Pile of Silky Sand”, and have Geodes drop instead. Then, since the Geode is no longer an item, you can scale up the Geode rewards for events + story / prices for items, and everything in Dry Top now actively rewards you even when you don’t loot a precursor (unlikely). I think that, as an example, would be the best way to showcase how loot as a general rule of thumb should work.

I also agree with Becka Williams.4978 above. I’ve been doing Triple Trouble for months upon months, wanting a Regurgitated Armor chest. However, I’ve never gotten anything other than the 70 Dragonite Ore reward. It’s hopelessly demoralizing and makes every reward feel like garbage because “Oh, just another day where I won’t get the only thing I want…” A token system where you received a token for every time you didn’t get an armor chest, and then you could buy an armor chest for ~20 tokens would alleviate that issue a ton, and make me feel like actively doing that event again to make progress towards a goal.

I imagine a modified random number generator that runs with increasing hit probability for every miss. You start with a probability, and every miss increases the probability for a certain event. A hit resets the probability to the start value. The probability goes up to 100%, which means a guarantied hit. The parameters you need are the starting probability and a value that defines how much the probability raises for every miss.

The probability should not raise linear. The start could be slow and a fair increase in the middle. It could be some formula that determines the raise depending on the current probability. Perhaps y=a*x^2. The coefficient a of that formula could be additional parameters that define the number generation of this event.
The starting probability is lower than the current true rng probability.

You have several events, for which the current probability has to be stored separately: a blue drop, a green drop, a rare drop, if loot drops at all, and such events.

To get to the new generator from the current one, the parameters should be tuned in a way that 50% is reached if 50% is reached by the combined probability of the true random generator after x tries, and 90% is reached after twice or thrice the tries for 50%.

This way, the first hit is a bit later than with the true rng. After some misses the probability is near the true rng probability, but if you didn’t get a hit after the intended number of tries, the probability finally raises so high that you must get a hit after too many more tries.

John Smith.4610:

munkiman.3068:

phys.7689:

Brother Grimm.5176:

…
I’m pretty sure that nobody think there is special code about that.

But we’re a lot to rather think that RNG is bugged in some way, and that’s why, in addition to a “gauss distribution” as described in the previous post, some players are always lucky and some are always unlucky. As the RNG mechanism is all in all quite complex (modified by magic find ; depends on damage done by the player / by his group compared to damage done by other players ; special items with a different table loot ; …), there could be several bugs introduced in the system. Remember that the % of magic find displayed in the hero panel is wrong since months (but the real number is supposed to be correctly applyed), so it wouldn’t be surprising that there are some bugs in the loot mechanism itself.

There are CERTAINLY very vocal players that are convinced of the “secret LUCK stat”.

its not really relevant weather there is a secret luck stat or not. If a random works properly, there will be people who will, will seem lucky, or those who seem unlucky, when compared to the data. This is the normal expectation of the bell curve.

So whether it is a bug or not doesnt matter. The point is these people are bound to exist.

My take on RNG, isn’t that it exist or it’s broken, it’s just that it’s practically everywhere. Neither of those solutions would work anyway, since the game would have to predict what you’re after in the first place, then reward you for failure of what exactly? Do we put a merchant in the game that carries everything a player might want in exchange for a loser token?

This was the first problem I thought of as well. It’s simple to say failure and success, but that’s not an easily definable concept in real life.

Two specific cases pop to my mind that really baffles me, ever since the implementation.
Fractal weapon drops, as well as Armor(both types)/ Weapon boxes.
Even if you get lucky, there’s a giant chance you get the shaft. And you’re left unhappy with an ascended item that you didn’t want. Or with a ‘rare’ drop that is absolutely useless to you (either you have it already, or you don’t want it).

It gets even worse with fractal skins. Seriously, I don’t want the same skin 5 times.
Same goes for account bound recipes, but they’re less of a priority I feel like.

Also, off topic, it feels more like GW2 is using RNG as a religion.
Every time a new desirable thing comes out we have to have faith in the all mighty RNGesus and hope we get it. We can’t progress or work for it. We have to sit and pray that one day it could drop.

i have no proplem with looting system ifen if it gives me grap for full week but only ting i dont like is the chance of mystic toild becous thos people who play all month tp can but price of 1.3k g and to only 2h work on tp not ifen dungon get all the gold and make econemy more up thad why i was hopying the precuser crafting coming soner soo it can set price range of 500-700g max what is 3month work for newer players

Silmar Alech.4305:

I imagine a modified random number generator that runs with increasing hit probability for every miss. You start with a probability, and every miss increases the probability for a certain event. A hit resets the probability to the start value.

I believe this is referred to as “weighted RNG”, and it’s a system that I know a lot of games use for loot drops, particularly for ones related to quest goals (WoW started implementing it for all of their “get 25 pristine pelt” quests, for example). I could see a system like this for lower tiers of loot, but I still would want the rarer stuff to remain un-weighted.

tonny.7580:

i have no proplem with looting system ifen if it gives me grap for full week but only ting i dont like is the chance of mystic toild becous thos people who play all month tp can but price of 1.3k g and to only 2h work on tp not ifen dungon get all the gold and make econemy more up thad why i was hopying the precuser crafting coming soner soo it can set price range of 500-700g max what is 3month work for newer players

You can’t expect trading not to be the most beneficial way of earning gold in any market where there’s an economy… There’s already fair mechanisms in the works to prevent too big margins but people know how to make money like that.

I don’t think it’s in an unhealthy state in game.

Inspired.6730:

John Smith.4610:

munkiman.3068:

phys.7689:

Brother Grimm.5176:

…
I’m pretty sure that nobody think there is special code about that.

But we’re a lot to rather think that RNG is bugged in some way, and that’s why, in addition to a “gauss distribution” as described in the previous post, some players are always lucky and some are always unlucky. As the RNG mechanism is all in all quite complex (modified by magic find ; depends on damage done by the player / by his group compared to damage done by other players ; special items with a different table loot ; …), there could be several bugs introduced in the system. Remember that the % of magic find displayed in the hero panel is wrong since months (but the real number is supposed to be correctly applyed), so it wouldn’t be surprising that there are some bugs in the loot mechanism itself.

There are CERTAINLY very vocal players that are convinced of the “secret LUCK stat”.

its not really relevant weather there is a secret luck stat or not. If a random works properly, there will be people who will, will seem lucky, or those who seem unlucky, when compared to the data. This is the normal expectation of the bell curve.

So whether it is a bug or not doesnt matter. The point is these people are bound to exist.

My take on RNG, isn’t that it exist or it’s broken, it’s just that it’s practically everywhere. Neither of those solutions would work anyway, since the game would have to predict what you’re after in the first place, then reward you for failure of what exactly? Do we put a merchant in the game that carries everything a player might want in exchange for a loser token?

This was the first problem I thought of as well. It’s simple to say failure and success, but that’s not an easily definable concept in real life.

If my understanding is correct, as far as the loot tables are concerned precursors are just named exotics. Thus, as far as the RNG is concerned you “win” if you get a named exotic whether it’s a precursor or not.

The reason why getting precursors are the most frequent example of RNG hate used is because players don’t equate other named exotics with precursors. Just as any precursor as a reward is far different than the precursor you want, or at least a valuable one. For example, I expect there would be quite a bit of rage if it were announced that everyone who hadn’t received a precursor in 3000 hours of playing would be sent a Rage via the mail to offset their bad RNG luck.

I don’t think it should work like that, any new system would have to start in the world in which we currently live. AND… any sort of token system should be used EXCLUSIVELY for the attainment of precursor weapons and t6 mats. So you get a stack of 100 bad luck tokens, you throw that into the mystic forge with exotic weapons and you get back a precursor that is account bound. If you throw it in with a small handful of t5 mats, you get a decent number of t6 mats in return. This rewards terrible luck over time and lots of actual gameplay.

Other than that, I think just setting each end of the spectrum to provide timed buffs/debuffs to your account when you have luck that is too good or too bad over a long amount of play would solve the issue. Call it Zommoros Fortune / Zommoros Curse.

John Smith.4610:

munkiman.3068:

phys.7689:

Brother Grimm.5176:

…
I’m pretty sure that nobody think there is special code about that.

But we’re a lot to rather think that RNG is bugged in some way, and that’s why, in addition to a “gauss distribution” as described in the previous post, some players are always lucky and some are always unlucky. As the RNG mechanism is all in all quite complex (modified by magic find ; depends on damage done by the player / by his group compared to damage done by other players ; special items with a different table loot ; …), there could be several bugs introduced in the system. Remember that the % of magic find displayed in the hero panel is wrong since months (but the real number is supposed to be correctly applyed), so it wouldn’t be surprising that there are some bugs in the loot mechanism itself.

There are CERTAINLY very vocal players that are convinced of the “secret LUCK stat”.

its not really relevant weather there is a secret luck stat or not. If a random works properly, there will be people who will, will seem lucky, or those who seem unlucky, when compared to the data. This is the normal expectation of the bell curve.

So whether it is a bug or not doesnt matter. The point is these people are bound to exist.

My take on RNG, isn’t that it exist or it’s broken, it’s just that it’s practically everywhere. Neither of those solutions would work anyway, since the game would have to predict what you’re after in the first place, then reward you for failure of what exactly? Do we put a merchant in the game that carries everything a player might want in exchange for a loser token?

This was the first problem I thought of as well. It’s simple to say failure and success, but that’s not an easily definable concept in real life.

There is the added problem of basically rewarding farmers, not that there are a ton left. I still think the broader brush of what really frustrates people about RNG, is an easier topic to handle. How to change it to make sure everyone has a fair shake only exacerbates the basic problems people have with RNG. I know we are looking for ideas to make RNG better/ more equalizing, but like phys said, it’s always going to have outliers. Could just be karma… I mean i did get a perma hairstyle kit, back in the day. If i knew it’s be worth almost a legendary now, i would have held on to it.

eekzie.5640:

John Smith.4610:

munkiman.3068:

phys.7689:

Brother Grimm.5176:

…
I’m pretty sure that nobody think there is special code about that.

But we’re a lot to rather think that RNG is bugged in some way, and that’s why, in addition to a “gauss distribution” as described in the previous post, some players are always lucky and some are always unlucky. As the RNG mechanism is all in all quite complex (modified by magic find ; depends on damage done by the player / by his group compared to damage done by other players ; special items with a different table loot ; …), there could be several bugs introduced in the system. Remember that the % of magic find displayed in the hero panel is wrong since months (but the real number is supposed to be correctly applyed), so it wouldn’t be surprising that there are some bugs in the loot mechanism itself.

There are CERTAINLY very vocal players that are convinced of the “secret LUCK stat”.

its not really relevant weather there is a secret luck stat or not. If a random works properly, there will be people who will, will seem lucky, or those who seem unlucky, when compared to the data. This is the normal expectation of the bell curve.

So whether it is a bug or not doesnt matter. The point is these people are bound to exist.

My take on RNG, isn’t that it exist or it’s broken, it’s just that it’s practically everywhere. Neither of those solutions would work anyway, since the game would have to predict what you’re after in the first place, then reward you for failure of what exactly? Do we put a merchant in the game that carries everything a player might want in exchange for a loser token?

This was the first problem I thought of as well. It’s simple to say failure and success, but that’s not an easily definable concept in real life.

Two specific cases pop to my mind that really baffles me, ever since the implementation.
Fractal weapon drops, as well as Armor(both types)/ Weapon boxes.
Even if you get lucky, there’s a giant chance you get the shaft. And you’re left unhappy with an ascended item that you didn’t want. Or with a ‘rare’ drop that is absolutely useless to you (either you have it already, or you don’t want it).

It gets even worse with fractal skins. Seriously, I don’t want the same skin 5 times.
Same goes for account bound recipes, but they’re less of a priority I feel like.

Also, off topic, it feels more like GW2 is using RNG as a religion.
Every time a new desirable thing comes out we have to have faith in the all mighty RNGesus and hope we get it. We can’t progress or work for it. We have to sit and pray that one day it could drop.

That was pretty much my point too, especially when you tie it into purchases. Live and learn i suppose.

Ayrilana.1396:

phys.7689:

Ayrilana.1396:

phys.7689:

Ayrilana.1396:

phys.7689:

Ayrilana.1396:

Essence Snow.3194:

Ok then……why isn’t it true, give some examples….Saying no without adding anything doesn’t help anything.

Just look at whatever items in the game have low supply but high demand. Look at precursors. Whether something will have a low supply and high demand is not in the player’s control but with Anet. They have their own vision on what they would like for each item. Since precursor drop rates are very low, and they’re highly sought after, it’s pretty safe to assume that they intended for this imbalance between supply and demand.

you can simulate the exact same output from random through other means. in large numbers random is fairly predictable and known.
for example, it is highly likely their is a consistent number for the current random implementation that you can represent like this.
X amount of enemies killed per precursor drop.
which also translates into X amount of player hours per precursor drop.

once you realize this formula, it wouldnt be difficult to create a set of conditions that can simulate the exact same output.

random is a design choice, its not a necessity in terms of predetermining supply.

I think you quoted the wrong post of mine as your post doesn’t have anything to do with what you quoted of mine. I’ll go ahead and answer you anyway.

You’re assuming the determination of the drop rate is based on player action rather than a manually added percentage. I may have read your post wrong. If i did, could you clarify.

Point is, anet can generate the same supply without random, if they choose to do so.

The advantages of random are not really about controlling supply, its mostly the benefits of unpredictability

I brought that up in a previous post. Supply of an item will increase greater under a system with a guaranteed than one on RNG. When players have a guaranteed way, they will flock to that way. Those with a lot of free time will farm that way nonstop flooding the market. With an RNG system, this would not happen.

You are oversimplifying supply. You can design.systems that produce the same behavior.
If precursor from mob drop is .00001 chance you can make precursors drop from killing 100,000 monsters.
If the average player hours in combat is 5000, you can make it take 10,000 runs of a half hour activity

If 200 precursors are generated per day, you can create content that awards to the top 200 scorers (limit pet account per month)

Point is for any randomized supply there exists a direct predictable method to provide the same supply with the same input

First off, something having a 0.0001% does not mean you’ll get it after 100,000 monster kills. If you were to set it exactly at 100,000 monster kills, you’d be cutting off the entire right half of the distribution curve and shifting those that would fall under it to the middle. This produces more supply as now those people acquire the items sooner.

It’s for that very reason that you’d be eliminating the probability of getting an item above the average which is why supply would increase more.

First I’m simplifying for the sake of clarity

Second
No it wouldn’t be cutting the curve because the on average the behavior is accurate. Supply is decided by the total over time which is predictable to a small %.
If the average is truely 1/10 for example, you will get 100,000 out of one million attempts. That 100,000 includes both sides of the curve. To produce the same supply in another system you might make it every tenth attempt.

If you are combining both methods you come up with a different equation.

Point is it just a case of mathematics you can produce the same results over large numbers with. Non random system as a random one. The benefits if random have to do with psychology/game design not output.

phys.7689:

John Smith.4610:

munkiman.3068:

phys.7689:

Brother Grimm.5176:

…
I’m pretty sure that nobody think there is special code about that.

But we’re a lot to rather think that RNG is bugged in some way, and that’s why, in addition to a “gauss distribution” as described in the previous post, some players are always lucky and some are always unlucky. As the RNG mechanism is all in all quite complex (modified by magic find ; depends on damage done by the player / by his group compared to damage done by other players ; special items with a different table loot ; …), there could be several bugs introduced in the system. Remember that the % of magic find displayed in the hero panel is wrong since months (but the real number is supposed to be correctly applyed), so it wouldn’t be surprising that there are some bugs in the loot mechanism itself.

There are CERTAINLY very vocal players that are convinced of the “secret LUCK stat”.

its not really relevant weather there is a secret luck stat or not. If a random works properly, there will be people who will, will seem lucky, or those who seem unlucky, when compared to the data. This is the normal expectation of the bell curve.

So whether it is a bug or not doesnt matter. The point is these people are bound to exist.

My take on RNG, isn’t that it exist or it’s broken, it’s just that it’s practically everywhere. Neither of those solutions would work anyway, since the game would have to predict what you’re after in the first place, then reward you for failure of what exactly? Do we put a merchant in the game that carries everything a player might want in exchange for a loser token?

This was the first problem I thought of as well. It’s simple to say failure and success, but that’s not an easily definable concept in real life.

The loser token is a simple solution,

You can create a hunter system, whereby drop.rate of certain items of which you select one, have the behavior of increasing chance/streak breaking

You can make it so the mystic forge has methods of altering drops (forge can change weapon types, stat distributions for example)

There are a number of ways to skin a cat. However I think the best way is to design it into the game design of rare items that their is a more predictable alternate method of obtaining them

Don’t we have that with the mystical magic find boosts? I REALLY wish arena would talk about what that does exactly, since it doesn’t seem to do much.

Being an outsider—#2 all the way. Exotics??? rare. Precursor—what’s that? Also include a check for those that lucked out and have a few, it will be a while until you get another one.

I would prefer to see both systems in action, perhaps in different areas.

1. For “boss specific” drops, like like the trinkets specific to an single boss, or the ascended armor/weapons that drop from Teq and Wurm, there should be both RNG and tokens. Every time you beat that boss, you are guaranteed to receive at least one token, then a decent RNG chance of maybe a few more, then also a low RNG change of just receiving cool unique loot. If you receive the cool thing you wanted via RNG? Great for you! If you completed the event ten, fifty, a hundred times, whatever is considered a reasonable spread, and still never got the rare thing you wanted? That kinda sucks, but you’ve accumulated a bunch of tokens, and you can go to a vendor someplace and spend those tokens to buy the rare thing you wanted off of him. Good luck can make the process smoother, but bad luck never makes it impossible.

Just as each dungeon has its own vendor, each boss with fancy loot should have its own vendor. There is usually a karma vendor near each boss anyways, it shouldn’t be hard to give them a new tab with these special loots on it. Also, ideally these boss tokens would go into the wallet like dungeon tokens, but if there’s some limitation on doing that, the Materials tab of the bank would work just as well. Just don’t make them “general inventory” items that would end up hogging space.

2. For certain systems, I would like to see RNG safetynets. As I described in the other thread from Marvel Heroes, their Eternity Splinters are designed to be RNG-based, but also guaranteed to drop every eight minutes if they haven’t already. Sometimes you can get several within an eight minute period if you have good luck, you might make 8-10 of them in an hour (on rare occasions much more), but you’ll never make less than 7.

Certain specific high-value items should be tracked in this manner within GW2. Definitely Precursors, maybe also Unidentified Fossils in Dry Top, maybe certain rare mats like Charged Lodestones. Maybe even have something that racks “have you found any exotic lately” and “have you found any ascended gear box lately” when doing activities that can potentially reward those.

I think the ideal thing to track would be a “per potential drop” counter. That is, every time you open a buried chest in DT, it has a chance of dropping a Fossil. Well each time you open a chest, and fail to get one, increment a count. Then either make it so that when that count hits a maximum, you automatically get a fossil the next pull, or just make it so that the higher the count gets, the better your odds get. In either case it resets when you get the item.

I imagine tracking too many different counts would put a strain on the system, so I would only expect it in a few specific cases of very low odds, certainly nothing like “have you received a Silk Scrap lately?”. Likewise, with Precursors, they have a tiny chance of dropping from just about anything, right? So just have a count that does up every time you loot anything or open anything and don’t get one, and whatever the average is meant to be, like 1:100,000 or whatever, if you hit that number then you get one, no questions asked, and the count resets.

And of course all that aside we still need “Precursor crafting,” some form of direct, intentional, no RNG progress towards getting a Precursor of your choice, account bound if necessary to prevent resale, and perhaps with limitations on how often you can complete the process.