Guild Wars 2

Holy crapper this is awesome!

I can’t imagine you invested so much work into this computer model. I really appreciate the effort because I’m much in favor of a more flexible scoring system.

However I’m not sure if I understood all the details of the model or the exact formula you were applying.

My concern however is that a very vocal part of this community believes that the scoring system doesn’t need fixing. They will probably be too close-minded to even consider the benefits of your proposal which is a real shame considering the effort you obviously invested.

Anyway, you got my +1

PS: Closer matches benefit everyone. Just look at the current Tier 1 battle which is close as hell. It’s way more fun than being the dominating force after the first night.

bump for a good idea, also it will never happen… the majority of the teams dominating other teams near the middle/lower brackets suck with even numbers.

Any multipliers based on population are a bad idea, you’re forgetting what the purpose of the score is.

The score is not primarily to determine a winner. There is no prize for first place.

The score is to determine a ranking to match you up against similar servers. Granted, this is currently compromised by the free server transfers, but never-the-less it’s what it’s for.

If you introduce a multiplier on low population servers to “even” things out, then you’ll end up matched against higher servers than you should be and you’ll be in a worse situation than you were otherwise.

thats a fair point Nabrok, but if you read what I posted you would understand the multiplier does not help low pop servers win. I know its a bit long, but ill tl’dr that part here: Basically because the lower population servers have trouble taking and holding objectives multipling their score for the stuff they currently hold has a negligable effect on the overall score. Compare the Adjusted and Current pie charts for the green team, the difference with the multiplier applied is only ~296.

Negligeable.

Part of the problem of the ranking system is when a server gets 500k points it goes up 2-4 placings and then gets dominated next week because of the snowballing effect of the score right now. The multiplier curbs that quite a bit which will prevent a server from jumping up so fast only to get knocked down quickly in the next week just because their opponents didnt show up.

The muliplier curbs the top end, but doesnt help the lower population win at all.

Also the muliplier stays in to keep the bonus’s in check. So if you cap everything when the other server is mostly offline you wont get a huge number of bonus points for it.

This actually looks really great, I would love to see this implemented.

Nabrok.9023:

The score is not primarily to determine a winner. There is no prize for first place.

The score is to determine a ranking to match you up against similar servers. Granted, this is currently compromised by the free server transfers, but never-the-less it’s what it’s for.

I don’t buy that. According to me and most of the community the score defines the winner. The current problem however is that a server can easily come out with twice the points of the 2nd server without actually being “twice as good”. A slight margin of player activity during certain hours in enough to secure a great lead.

Yet the actual performance doesn’t scale with the score. The score scales exponentially.

This means the leading server will just find itself in a matchup completely outclassed again.

Currently a 100% lead on another server only means you were “slightly” better or had just a few more people on every night.

Nabrok.9023:

Any multipliers based on population are a bad idea, you’re forgetting what the purpose of the score is.

The score is not primarily to determine a winner. There is no prize for first place.

The score is to determine a ranking to match you up against similar servers. Granted, this is currently compromised by the free server transfers, but never-the-less it’s what it’s for.

That’s like saying that a breathalyzer isn’t to determine who’s drunk, it’s to determine who’ll spend the night in the drunk tank.

Seems like a complicated system to explain to players. Anet could just limit a server’s ability to out man the other servers with population controls on the maps.

Aurainsoph, I do not think thats an option.

You should explain why that is not an option.

rook.4625:

Regardless of the score deciding where servers rank, closer scores keep players playing, period.

+1

Also consider the following:

Its thursday, scores are near tied. But your fort has been capped for over 3 days and so has their garrison on their BL. They are both fully upgraded.

If either server manages to cap the other servers garision while pops are competative it might give enough points for your server to win the matchup.

Puts lot more volitility into the scores near the end, and I think the better server will win.

Aurainsoph.5426:

You should explain why that is not an option.

It’s the general feeling i get when I read Witters post. He did post that improving the scoring system is on the table, and didnt feel … well just read it.

We feel that everyone should have the ability to contribute no matter what time of the day it is. How we’ve came to this conclusion is that no player’s time is more valuable than another. Everyone has different off peak hours for whatever reason. Players should not be punished or unable to experience and view the same content as everyone else because they play at a different time. They too are paying customers.
This is not saying that we are against any adjustment to scoring, or against developing another mechanic to improve the capping system. This is simply saying that we will not be changing WvW based on some players’ idea of off time hours.

> Assumption 3: You need a minimum of 2 people to take a supply camp

i don’t really agree with this…

i can take an unupgraded supply camp as a single player by just isolating 1-2 enemies, have a bear tank and justPvE tank and Spank. minimum required is 1 but the time required to take a camp is negatively exponential.

Populations unbalances isn’t just an off peak hours issue, some servers will always out number other servers they play during all hours.

Luriyu.6873:

> Assumption 3: You need a minimum of 2 people to take a supply camp

i don’t really agree with this…

i can take an unupgraded supply camp as a single player by just isolating 1-2 enemies, have a bear tank and justPvE tank and Spank. minimum required is 1 but the time required to take a camp is negatively exponential.

I am aware you can take it with 1. But you would have to account for the fact not ‘everyone’ can do this and its likely when only 1 person is on the map that they wont be taking a supply camp.

Often I dont see supply camps flip on a map until there are at least 5. But because that assumption is applied to all 3 servers in the data, it doesnt really affect the outcome of the model.

Same thing with towers and keeps, you CAN do it with less people, its just unlikely that less total people will be on the map when you do it.

For example you attack a keep with 7 people and take it, but while you were doing that there were 3-10 randoms running around on the map for your team.

The problem with WvWvW isn’t the scoring system, the problem is that winners win more. As a strong server captures more and more stuff, it becomes easier for them to steamroll their opponents.

Regardless of what the score is, if one server owns 95% of the map and all they need to do is play whack-a-mole it is depressing and makes me not want to play. Furthermore it also kind of breaks your “bonus” system since shortly after losing teams recapture something it just gets taken back again by overwhelming numbers. In practice, losing teams do not hold onto stuff for very long, only winning teams. I suggest you take your algorithms into question.

Here is how you balance population differences:

Why is it that the WINNING team gets +15% health and 150 all stats from orb bonuses, while the LOSING teams get extra xp / mf / karma from outmanned? All this does is allow the winning teams to pull further ahead, making for steamrolls when we want close games.

My solution is to give players increased damage vs players/gates/guards and reduced damage from players/gates/guards. This bonus should be directly proportional to the current population difference. If your server has 1/5 the players in WvWvW to another, you should be able to fight players 1v5 and have it be a roughly equal fight. If for some crazy reason you are the only person online on your server, you should be able to solo keeps or even stonemist if your opponents don’t make a decent effort to defend.

You have a really good scoring model here Draygo. Perhaps even better than the one Anet has currently for wvw. Unfortunately you can’t factor player morale into any equation. I think that even in the way off chance that Anet did implement this, the scores would still be one sided because the majority of the player population gives up way too easily in wvw.

In my experiences, I have been in situations both where I am on the spawn trapping side and on the spawn trapped side. So many times when I am on the spawn trapping side the side that is spawn trapped actually has way greater numbers than the people trapping them there. They don’t seem to want to win as badly as the winning team. Now, think now about the times in military history where an army backed into a corner facing overwhelming odds actually snatched victory from the jaws of defeat. The opposite happens in a video game it seems, because retreat is still always a couple clicks away. Not to mention there is no real fear of death, and these days it’s too easy to defect to the winning side anyways. Other than bragging rights and pride when your server actually does win there is no incentive for players to stay with the losing servers. I think we can all agree that free transfers need to end.

Also, I’m not sure how I feel about the losing teams getting buffed this dramatically. Whether or not a wvw match up is won in the first few days, and match length in general, are other matters for debate. But if that is the case then tough luck, come back and do better next match. It sucks losing, but not everyone can win either. That’s what makes winning feel so good. And frankly, when I am winning I don’t want to know my work is being rewarded any less.

Just look at the tier one match up this week. In spite of all the problems with wvw in its current state (culling, free transfers, night capping, hacking, etc.) it’s still a very close match up and everyone seems to be enjoying themselves. I think it can be attributed to high morale on all sides. Although I do think the scoring system could use some minor tweaks, that match up seems to be evidence to the contrary: That there is nothing wrong with the scoring system, and the player morale is what causes losing teams to be blown out to begin with.

The scoring system I proposed here would not affect even matchups like whats going on in teir 1 right now. It will add some tense and wild things by making objectives that have been held for a long time worth points to capture, so you will see teams defending and attempting to take fortified keeps up until the last moments.

Anet is looking to address the camping situation in a different way to allow camped teams to push out. And score does affect morale, my goal is to prevent situations where people simply get depressed when they look at the score thinking they cannot do it or cannot make up the difference, that its pointless.

The system I am proposing here lets comebacks happen, doesnt let servers sit in lame duck or kingmaker situations that are poison to a good system.

So, in your opinion, how do you think this would affect a matchup that looks like this?

http://puu.sh/1hm1h

I doubt anything, other than flooding the other two servers with players, would solve this, but I am interested in anything that might make it slightly better.
Similar thing happened last week. It’s not very fun.

Your bonus system is a huge snowballing mechanic, and in fact a worse one than orb and downed state combined.

I have no idea how you model all the details like population fluctuation, time period, etc, but from the close and small final numbers, I’m going to guess that you model the tower and keep to change hands more often that is currently the case in WvWvW.

I’ll just illustrate the snowball happening here by taking EB as example. In fact, I’ll only look at the 3 keeps. Despite lacking in NA prime time presence (but not sorely lacking), server 2 should be able to hold its keep at all times. What’s the bonus score after a single day? It’s at least 20*24 = 480 points, 2 days and its at least 960 points. Now during the first oceanic prime time, server 2 will be able to cap all other keeps, thus resetting the bonus points for all other servers to 0. So by the second day, server 2 will be enjoying a 480 bonus points on its keep while every other server’s bonuses are small. It gets worse and worse as the match progresses.

So, unless you publish in more details your model, I have a hard time buying your proposal.

Don’t get me wrong, as I stated at the beginning of my post I think you do have a really scoring system put together. But there are some pretty radical things proposed in it that would still affect evenly matched games. For example, your scoring system rewarding teams more for capturing objectives would make for juicy targets late in a close game. Transversely, it would also make recently lost objectives less attractive targets, which would make the better choices of where to attack also the more predictable choices. I’m not sure if this is a good thing, as I haven’t seen it in practice, but I definitely get the impression that it would make things odd, to say the least.

Also, as someone who is clearly good with numbers, I would expect you to know that even in your system there will come a point in every match up where winning is mathematically impossible for the losing team. And while these numbers are unclear to me as I don’t have the exact equations for calculating scores in either your system or Anet’s on hand, you don’t have to be mathematically sound to get a sense that you are at that point as the losing team. Although your system would increase morale by keeping the scores closer, your assumptions are also assuming that the losing team is being uncontested while taking a supply camp, tower, keep, etc.

There is also the problem of teams that may intentionally play from behind, or find some other way to exploit the scoring. History has proven that gamers will always find a way to exploit a given system. Like the once great Mike Tyson said, “Everyone has a plan until they get punched in the face.” If implemented, your system would fix a lot of a problems, but it would cause a lot of problems as well. The real question is would it fix more than it solved.

Correct Thievery.

I am having a hard time trying to find exploits in the system. Assume worst case where a team just gives up all its points and doesnt participate

and the other two servers somehow got into an even fight where none of them took anything.

Then on the last day thursday that server took everything everywhere.

In that case they get 80%ish of the total points. I really dont see that happening. Because holding points is always worth more than taking and leaving, its not to the advantage of a team to take and leave.

Also would juicy targets in a close game be a bad thing? If you are losing going all out for a hard to take well fortified objective might be the gambit you need, you can also fail flatly to take it.

Right now the model is using a 2% chance to take a fort in a 15 minute period if pops are even. Chance goes up the larger the population gap. Lowers if you outpop. its something I can configure on the fly, and i played with it a bit to get something that seemed realistic.

@Newbienince the score model i’m using has one server take over almost all the map in oceanic time each day. Let me see if i can pie chart that.

Below is the snapshot at +18 hours in the model.

If you want the exact details on the model, pm me and ill see a if I can get it to you.

The biggest problem I see is that there will be a big rush in the morning/afternoon, and whomever has the best Europeans will start winning in the same way that Oceanics are now, point-wise.

How so? Remember its based on population difference not time slot as far as the scoring goes.

You rush europeans when no one is on you wont get as many points if you took something during primetime.

The bonuses that keeps accrue the longer they’re held. You see, Maguuma, my server, has a large European presence. So we dominate pretty much 95% between noon and 3 PM EST. Our Europeans would get all of the bonus points that the Oceanic players had built up between 5 AM and noon EST, giving us huge bonuses.

Not if the bonus is reduced 90% because you are PvD’ing.

For example the Oceanic server is capping everything during oceanic prime, and isnt getting many points off it. You will get probably ~200-300 more points per day. I can rerun the model on a euro timezone advantage if you like.

Your model seems to say that oceanic caps are less valid than prime time caps, is that your intention?

no
It says that PvD is worth less than PvP

If an oceanic server fought an oceanic server, the scoreing would be equivilant to prime.

Okay so you only get the points if you’re actually playing against a similar population, no matter what time of day it is? Sorry I found your post a bit wordy and wasn’t entirely sure.

The problem with your model has already been pointed out. ANet has clearly stated that no one group of people’s time is to be worth less than another simply becuase their prime time is different. Your model essentially makes Oceanics worth less than prime time players, which won’t fly.

The current scoring system is fine, orb and outmanned could be switched but its not that bad; the broken part is free transfers. In more than one situation the population for WvW was spread nicely, and due to free transfers some servers got stacked and some servers became ghost towns which is what is really ruining matchups; and no scoring system is going to fix that. They need to announce the close of free transfers, at that point people will make their final choice, and the population spread out some. And if that doesn’t work they could offer incentives to transfer to lower rank servers.

TL;DR Free transfers are the problem not scoring, these transfers are whats imbalancing coverages and causing 90% of the current issues. Thats what needs to be solved NOT scoring

The point is obvious with the Oceanic players’ time not being worth the same as American players’ time. This goes against what A-Net has said regarding night capping.

Even as it stands Night capping isn’t an issue. The issue is that servers consistently win WvW matchups by more than double the points of the other two servers combined. And this model does nothing to address that.

Props though on doing all that numbers work and research. This is very well thought out, and it’s awesome to see community members going all out like this.

I wholeheartedly agree with the op. This is by far the best idea to fix wvw I have read so far.

To those who are saying there is a time value issue, Draygo has accounted for that in his model. It’s in the second post.

@Draygo
It wasn’t clear but is there a cap to the bonus points for holding key areas? If there isn’t i have to agree with some of the posters that it can get ugly once things start snowballing.

I really liked your model Draygo. Everybody makes some comment in either positive or negative way, but we can’t know the exact result without trying this. I personally don’t play WvW when my team is heavily outmanned that week. I think your score model can bring solution to this and give me the fun i miss in WvW.

“the longer they remain uncaptured by an enemy team per 15 minute interval the more bonus points they are worth.”
Also, I loved this idea, because it will completely remove the zerg mentality and people will start to defend their towers and keeps in my opinion. Currently people doesn’t care about losing a tower, they say “never mind, we can capture it again”.

+1

What I’m saying is that the bonus points accumulated per tick after one hour would heavily favor the primetime immediately after Oceanic: European.

Look at the Darkhaven/Maguuma/NSP battle right now. Darkhaven totally controls all three maps for about 6-8 hours (normally starting around 5 AM EST). That means that the next prime time (which would be Europeans during the afternoon) will get all of those bonus points. Maguuma always controls about 75%-90% of the map around 2 PM EST. So, they get all of the juicy bonus points that Darkhaven has accrued overnight. Which, by your system, is around 2000 bonus points every day. That eventually accounts for over 14k more points for Maguuma.

I’m not saying it’s a bad system, but it unfortunately heavily favors servers with a European presence since they will probably retake the maps entirely, and it will be less risky for them, because they will hold those points for less.

Unless I’m misunderstanding you, and servers get more bonus points for holding something longer, not making their keeps worth more? If that’s the case, then you’re still rewarding night capping.

I’m not against this idea… but I see a few holes in it. If I’m misunderstanding, I’m sorry. I can tell you put a lot of work into this idea.

>The rules of the bonus system:
>After 1 hour for Forts and Towers and 30 minutes for camps, the longer they remain uncaptured by an enemy team per 15 minute interval the more bonus points they are worth.
>After 1 hour Forts accumulate 20 bonus points per tick
>Towers accumulate 8
>Camps accumulate 4.
>So it always a benifit to hold something as long as you can, as you get more points that way.
>The result?
>Here it is:
>Server1: 46055.5
>Server2: 50759.3
>Server3: 36372.9
>Bam, ‘night capping’ is eliminated. Server 1 is within 4k points for a single map at the end of the week, a catchable lead.

I may be reading this wrong but it looks like you’re saying that the longer you hold something the more points you get. Doesn’t that mean that the server that holds 100% of everything 24/7, which seems to be a popular theme, would exponentially get more and more points?

What I ment to address that you get more points for holding something than intentionally losing it, to recapture it later.

You get 25 points a 15 min interval for forts right now, the bonus system only racks up +20 per 15 min interval after 1 hour has passed.

That and the other server might take it before you do means its better to hold an objective than to let someone take it.

The bonus system is letting a server take something back that was capped overnight and recoup part of the points for it, preventing the score from snowballing because a server has a population advantage at particular hours (not just oceanic).

Basically if you want to move ahead in score under this system you have to do better than average at all hours, including night hours than your population is expected to do. Servers with oceanic guilds still have large advantages, this aims to tampen down the score disparity between competing servers.

Also Oceanics are not worth less. I am just using an oceanic timezone advantage in the example, but essentially if your server at any time is PvD’ing your points are going to be worth less for that time period.
Plus things that are captured tend to stay captured if your server has good coverage.

Anet can easily tweak the system to put in caps on bonus point accumulation for example, or tweak any one of the variables to adjust how harsh this system applies its rules.

Oceanics are still a good thing to have, they still get you a lot of points, they just dont give you an explosive lead.

To the person saying that oceanics would be ‘worth less’. Thats hogwash. Currently they are worth much more than a primetime player, this brings their worth in line, makes it equal. And rewards effort against fighting advisories while still giving an advantage to a server with more people.

And yes sorry marm i should be more clear. Bonus is granted on capture to the capturing team.

So according to your model:

If another side is beating you, the best tactic is to log out, because then they gain less points?

• So in a few days you will see mapchat filled with randoms yelling in caps: LOG OUT YOU FFF YOU ARE BETTER USE WHEN NOT INGAME

• You can follow that up with: No point in going to WvW, Server X doesn’t have anyone online anyway so we can’t get score.

• Noone is allowed to go into WvW unless it’s Y o clock. Zoning in alone will hurt the server, etc etc.

You cannot fix this by making it unfair, you will just break it even more. – A little more though has to be put into it.

Ern if you can somehow get the other server in the matchup to also log out, that might be detrimental.

getting your own server to log out only hurts your server.

To illustrate: This is what would happen to the adjusted+bonus score if server 3 did what you said above:

It loses, badly, as expected.

What i’m saying is I already thought of that.

This is pretty amazing, I wish more developers posted things like this for me to play around with. Defiantly in favor of changing the current structure.

Just to be clear, I’m not a developer for arenanet.

Its late, ill try to answer more questions tomorrow, if you want a look at the raw data I’m working with, send me a PM and be availible to get on teamspeak.

I like some of your ideas, and definitely something has to be done about wvw scoring system, so +1

Alright then, so if your server has no players on, or few players on, they will lose badly, per your new graph – Earlier however you said that would not be the case?

You mean that not only do they gain less points when we are outmanned, they also crush us when we are outmanned?

I don’t get it.

Either you have an even (linear) curve, or you don’t. In an uneven curve, there will always be an equilibrium of perfection, where you have less people that them, and any added people to your zone will decrease your pointgain, or their pointloss, whatever way you wanna see it.

Ernest, thats what I thought at first too when I first did a multiplier only.

I found out that the multiplier itself had little effect on outcomes. Because when you have less people you have less stuff. You tend to take stuff back once you start fielding the population capable of taking things back.

You still have to field people, the idea is to normalize it a bit so that one servers population curve being different isnt the sole reason they win their matchup. They have to compete, or attempt to compete in all time slots. If your overall population is down, it is likely that you will lose the matchup. And thats just fine. This scoring system aims to prevent snowballing when one servers population curve is significantly different than the other servers in the matchup.

A well thought out and balanced system. It promotes more fights and causes the matches to be closer, which will overall cause the system to be used less.

Elegant. Well done.

Please, please implement something like this. The snowball effect of WvW scoring is really getting old.