Guild Wars 2

Faux Play.6104:

The calculation should be against your rating not the teams. However, there shouldn’t be huge variations in the rating of team members. The issue is with the matchmaking not glicko. Right now we are sacrificing short queue time for poor quality matches.

The matchmaker cannot possibly provide high quality matches with garbage ratings being fed into it. Ratings, and rating differences are supposed to represent likelihoods of outcomes, and when the matchmaker is paired with a consistent rating system the ratings will converge to make the predictor true.

But that doesn’t happen because the matchmaker and the rating systems are inconsistent.

For instance, the matchmaker has baked into it some logic that says that if you put high platinum player into a mid gold game (that is, a player with a +200 rating difference), that it would expect the platinum player’s team to win 57% of the time. That’s a bit lopsided, but not so unacceptable that it can’t fire the match off. Unfortunately, that assumption isn’t true – because the player rating update is inconsistent, the platinum player’s team actually has a 76% chance of winning the match.

What is the matchmaker supposed to do when it can’t accurately evaluate the quality of a match?

The matchmaker can only create fair matchups if it can accurately predict the outcomes. If the ratings being given to it don’t mean what it thinks it does, and update in a way inconsistent with those assumptions, it’s going to keep spitting out crappy matches with no correction mechanism.

Faux Play.6104:

Also, we really aren’t using glicko it is more ELO. This season they disabled the key difference between the two systems which is increasing rating deviation due to inactivity.

They aren’t using either system – again, Elo is a consistent system, while A.Net’s special snowflake rating system is not.

A.Net’s special system does not find, or try to find, fair match-ups for its top players – it instead finds weak opponents to feed to those players in blowout after blowout. That is probably not the design intent of their rating system, but it’s what theirs does.

Cynz.9437:

This is not anecdotical evidence. You should try to duoq with someone who is 100 or so ranks below you. Ask them how many ranks they gain/loss for every match with you. I promise you, it will be something like: you gain 10, they gain 20, you lose 20, they lose 15. The system currently does base rank gains/loss on personal MMR vs avg MMR of the enemy team. It does not take into account your teammates at all and given wonky MM that creates matches like plat/gold/gold/gold/gold vs legend/legend/plat/plat/plat on constant basis it becomes really punishing for no good reason.

Yep, then the rating and matchmaking system is simply wrong. Broken systems produce garbage, which it uses as future inputs in a long cycle of garbage.

A.Net should probably fix that before looking at anything more detailed. Hard to talk about improvements to Glicko when you don’t have Glicko working yet.

Animism.9803:

However, if Team A loses to Team B, the player with 1800 rating loses more points than that player with 1200 rating on his team.

Right, that is how it seems to work. That means that A.Net’s implementation of Glicko is mis-specified. It’s simply wrong – everyone on team B’s rating should move up or downward the same amount (with allowances for differences from rating volatility).

What they are doing, with the higher rated player losing more on losses than the lower rated player on the same team, is not Glicko or Glicko-2. It’s mis-specified. It makes everything about matchmaking and ratings inconsistent (matchmaking and match prediction is based upon rating numbers representing something specific, but they don’t actually represent what they are supposed to – because they are computed wrong).

It means players are rated incorrectly, which means matchmaking isn’t building teams appropriately, creating blowouts that it doesn’t think are blowouts – garbage in, garbage out.

Updating their MMR algorithm to actually be a correct implementation of Glicko-2, or some consistent derivative, should be really high priority. This should have been fixed years ago – there’s really no excuse for a high budget studio like A.Net not to run their MMR and matchmaking code past a statistician at some point.

phys.7689:

so essentially, players have been trained in gw2, that gold has little value. hoarding is logical because anet has built an economy where you will probably need many things, so there is no value in selling them, because you ll just have to buy it later.

I think this is touching on something important that is a little tricky to pin down.

If you look at what players get from drops, it’s this mix of materials that everyone gets a ton of – and, importantly, everyone gets roughly the same mix of. It’s not like some people do content that produces blood and another content that generates fangs and they need to trade to get the mix they want – with a handful of exceptions, regardless of where you play in the end game you get more or less the same stuff.

Demand similarly takes somewhat different ratios of that stuff, but still draws on big piles of the same stuff.

So the big flows of trade, that you need gold for, are shifting between the very rare drop / gem store skin markets and the commodity markets. You trade rare thing for rare thing, or rare thing for a big pile of common stuff, or a big pile of common stuff for other common stuff.

If you aren’t trading out of the common stuff to get a rare skin, there’s basically no reason to sell, since all the goods are generic. You’d just be selling to buy back later, holding wealth in gold in the interim.

Within that, it’s interesting, because you have a couple different mechanisms. Mystic Coins have always been pretty interesting since they are time gated and can be traded, so they act like a negative beta good – when other stuff is rare and expensive, coins are cheap, and when other stuff is cheap, coins are expensive. So you see coins marching upward, which implies people are flooded with all the other crap needed to make endgame crafts. So far, so good.

Hardened leather, well…if you look within the bundle of commodities, leather, metal, fine mats, etc, everyone has a steady income of those, and you can’t really do much to tweak the ratios other than hitting that centaur camp hard. In that case, you’d expect prices on the higher demand (relative to supply ratios) items to, over time, eventually encompass the whole value of the output. Hardened leather looks like that item, so it’ll keep rising in price (while the value of non-hardened leather commodities, that are produced as byproducts of hardened leather, drops proportionally).

The former I understand pretty well. The latter, well…it doesn’t have great real world analogues, but I think the case is pretty clear – it’s the high demand portion of a farming process with tons of waste products, and thus captures a disproportionate bulk of the value of that process.

Reikou.7068:

2) All MMR gained/lost from a match is calculated based off of average TEAM MMR, and not personal MMR.

I’m going to presume this is true since it’s consistent with anecdotal evidence, and if it’s true it is 100% a mistake. Like, your model is inconsistent and you fail your statistics exam kind of mistake.

It means your matchmaking is using a fallacious basis and your leaderboards don’t sort properly.

It would be good to get a confirmation to the contrary, because if the above is true GW2 isn’t using Glicko or Glicko-2 or anything of the sort, it’s using its own custom MMR that’s basically just wrong.

phys.7689:

its interesting because the implication is gold has very little value in this game. People would rather hold items than turn it into gold.

Well, gold is just a medium of exchange. You can hold gold to store value to spend later, or you can hold other assets. Other assets aren’t costless, but it’s cheap (storage space), and it does insulate you against demand surges – for instance, being able to make legendaries out of inventory instead of having to buy a bunch of materials off the trading post saves me both time (since consuming from inventory is instant) and money (since I don’t pay trading post fees on my inventories, let alone bid/ask spreads).

Plus there’s a hedge against inflation or surges in general – T6 is more volatile than gold, ectos track a bunch of markets, mystic coins are negative beta…

But the real question is how much do people hoard materials, as opposed to hoarding gold. For instance, the tidbit from A.Net that people ‘save’ 10% of net Mystic Coin generation (10% more generation than consumption is a 10% savings rate, net). What is that for other goods? What is that for gold – how have gold savings increased over time, even with its high churn and taxes?

voidvector.2780:

I am probably not going to be able to model it with such detail (both because of data available and my own time). Maybe someone working at Anet would be able to? And potentially blog about it?

Well this is less about what the actual data looks like and more an exercise in checking your own assumptions.

For instance, you might expect prices to keep rising if the gold-to-coin ratio is increasing over time. I’d push back and ask why that expectation isn’t priced into the market price (insufficient investor liquidity), but it’s not an unreasonable assumption. Ok, but what if that ratio is decreasing? Why would it decrease? Well, if people are hoarding coins but not hoarding gold, you’d expect that ratio to drop as gold is sank but coins aren’t.

voidvector.2780:

I have been thinking about pulling the API, and try to measure the inflation. Maybe I will look into it.

It’s not easy since we do not have readily available volume information, but would be really cool if it could be estimated consistently. Just know it’s not going to involve pulling some data points from the API, but doing a lot of monitoring studies.

voidvector.2780:

My argument approached it from a macro-economic and behavioral perspective.

If you want to approach it from a macro perspective, be very explicit about your assumptions about the relative growth rates of the three asset classes (gold, coins, hard leather) over time – both flows and stocks – as well as your assumptions about their respective elasticities and how demand changes over time.

Which set of assumptions is necessary for your model to hold?

I am pretty sure there is a set of assumptions that is consistent with your conclusions (when combined with a pretty high discount rate, which I think you can take as a given), but there isn’t a ton of wiggle room in those assumptions and they aren’t obviously true.

voidvector.2780:

Also when I said “end-of-life”, i am actually referring to time at which the game is near the end, but actual time can be better predicted with micro approaches.

The discount rate is high enough that I don’t think the end-of-life condition is going to be a big deal.

voidvector.2780:

If there is inflation, the inflation-measuring basket would outperform gold.

If inflation is positive. Is inflation positive? Why should we expect inflation to be positive?

voidvector.2780:

There might be short-term deflation, but I am doubtful there is long-term deflation since farms like Silverwaste/Fractal40 don’t suddenly become less productive.

From a boring monetarist perspective, inflation is positive if the net rate of currency generation (relative to the stock of currency) exceeds the net rate of non-currency asset generation (relative to the stock of assets).

Is Sliverwastes farming inflationary? That is, is the ratio of gold to assets that it generates higher than the ratio of gold to assets in the game as a whole? If it is lower, efficient Silverwastes farming would be deflationary, not inflationary.

Etc.

Outside the initial burn-in of the game in the first year when the money stock was developing, is there a good a priori reason to think that inflation is still positive?

That…no, that’s not how asset pricing works.

There isn’t any interest or rent producing capital in the game. Expectations about the price trajectories of various assets are priced into their market prices.

I.E., if it was certain Mystic Coins were going to appreciate in the future, people would start buying them now at higher prices – until the price is high enough that people didn’t see any more upside to holding coins over holding gold anymore.

I would argue that you’re cycles of falling risk premiums (as A.Net does nothing to address the fundamentals), followed by rising risk premiums (as prices spike and the risk of intervention rises).

Overall, though, no, there’s no guarantee that there are assets that will outperform gold in the long run. There’s no guarantee of inflation in the long run, and there have been plenty of short run periods of deflation.

Astralporing.1957:

Remember, that when Anet introduced this gold daily reward, and some people were protesting, the answer was that it’s a significant increase of income for a large part of the playerbase. Seriously, most players do not do dailies every day. Most players do not have daily average incomes counted in many multiplies of gold coins. For many players, 150g is a gold amount they never had at once – a fortune.

The flip side of that is that most players do not play the game every day. It’s not meaningful to talk about daily income when the hours played per day can vary immensely. What is relevant here is income per hour of play.

We have not-completely-unreasonable estimates of that from gw2efficiency – look at core cases (500-4000 hours played distribution and 15-85% wealth distributions within those buckets) and you have an average income in the 7-8 gold per hour range. Yes, this is of players engaged enough to be on gw2efficiency, but that’s your core player base 4 years into the game, yes?

The players playing less than an hour in a play session, only a couple times a week, are not players looking at buying up stacks of mystic coins to make legendaries.

A 2000 gold legendary represents the wealth from about 250 hours of gameplay at that median. At the median attached player rate of 1.3 hours per day (casuals will be even less, of course), you’re looking at around 200 days to make a legendary.

Are Mystic Coins really being added to the game too slowly for that demographic?

Swagger.1459:

2. Reduce many weapon and slot skill cool downs, because long timers are really anti-fast paced and twitch play. These simple changes would also be the answer to improving core professions.

Shorter cooldowns are generally more interesting game play but unfortunately GW2 has far too many cooldowns to make that happen.

In an ideal world you have far fewer skills and traits, all of them are high impact, and playing around their (shortish) cooldowns is where you get play. But in GW2 everyone has kittentons of skills and what often feels like an endless supply of cooldowns to burn through before you can accomplish anything. If you’re going to kill a dragonhunter you aren’t counting dodges like you could in the old days, you’re popping the traps, counting the blocks, waiting out the F3, waiting for it to refresh then waiting it out again, then maybe you can actually do something if you’ve put out enough pressure to get those all out before they start refreshing naturally.

You have just endless streams of nonsense to stay alive, but if you don’t have that or make use of it well you just pop from passive ticking AoEs and cleave that’s being puked out.

I’d like to have a conversation about cooldowns, but it just isn’t relevant in the current state of the game. There’s just too much of everything.

I would suggest taking a sledgehammer to the Gossamer Patch recipe before looking at any further changes.

This would be a perfectly fine effect if it were introduced alongside other very strong, strategy or tactic-specific sigils. It’s a perfectly reasonable design when there are important opportunity costs that you are giving up to get it (in the same way that traits or utility skills have really valuable opportunity costs when you select them).

Introducing it alongside a massive sigil nerf, on the other hand, is almost certainly poor design – without real opportunity costs, powerful, targeted counters just make the game a lot more random.

Unfortunately the design space of sigils at the moment is pretty thin. You essentially have five categories of sigils – permanent bonuses, on hit/crit, on kill, on swap, and on interrupt.

Of those, on kill have no place in PvP (I think you can remove them categorically from the PvP build panel with no real loss), and permanent bonuses are, ideally, a low power fall-back option if nothing else is effective – if permanent bonus sigils are seeing a ton of play that is indicative of design failure in the other sigils available.

On interrupt sigils are not on the PvP build panel, but should be. That’s a low hanging addition to be made.

On swap sigils are generally a good set of designs. Hydromancy and geomancy are the only real concerns here, since they are essentially zero cast time, pretty powerful attacks. In a PvP game about reading and reacting to animations, zero animation nukes (or CC in the case of hydromancy) are out of place. I think re-working hydromancy and geomancy to work the way leeching does, triggering on your next attack after swapping, would be a good step here – you’d still have access to those powerful effects, but there would be counterplay from dodging or blocking the next attack instead of getting instantly hit by it.

The on-hit and on-crit sigils should all have a 100% proc chance. There’s seriously no reason to layer RNG on top of RNG when it comes to these effects. To the extent that’s done simply to make the effects weaker, well, just make the effects weaker. Weaker, more consistent effects are more healthy for PvP.

On top of that, I’d argue against pure damage (or pure condition damage) procs. Those just further encourage burst builds and don’t add much in terms of gameplay. To the extent possible, procs should have both a damage and a utility component. For instance, I’d suggest lowering the damage on Sigil of Air by about a third, but having it apply 3 stacks of vulnerability for 5 seconds; similarly, adding a chunk of lightning damage to Sigil of Nullification to have it deal some damage and remove a boon. This flattens the opportunity costs (and let’s be real, for the most part people just pick the damage because the utility effects are so minor in comparison) and encourages choice for the secondary effects. It also gives you an additional knob to tune on each sigil, which is helpful for balancing them.

I’d also like to see the AoE sigils pushed a bit more – I think they are fundamentally healthier designs with better high moments (hitting multiple targets with a proc), but are generally woefully underpowered.

As for other sigils? Cleansing would be desirable. A boon steal on crit would be desirable if appropriately powered.

For more design space? There are some conditionals that I think are interesting. Effect vs moving (or not moving) foes. Vs foes using a skill (or not using a skill). Vs an incapacitated foe. Perhaps not stuff you need immediately, but to think about down the line.

In the short run, I’d stick to 4 steps:

- Remove on kill from sPvP
- Add on interrupt and cleansing / torment / blight to the PvP panel
- Re-work Hydromancy and Geomancy from instant effects to on next attack
- Start design re-works on on-proc sigils from high damage or effect to moderate damage + effect.

Without even looking at buffs and nerfs:

- The value of an interrupt is way up in HoT, there are a lot more crucial defensive skills that don’t have any recourse other than interrupts (and we’re no longer in a world where many of the best heals blocked/evaded/were passive).

- The value of an individual boon strip is way down – boons are spammed out in large numbers, so while pulling 1-2 boons used to be a big deal now you really want classes that rip half a dozen or more to get through all the crap.

- Immobilize has lost a ton of value. Before the 6/23 patch and HoT there were times where immobilize was higher value than a stun, as it could be harder to remove and would force big cooldowns to be blown. Now several classes have trivial answers to immobilize, area condi cleanse is much more prevalent, and there are short cooldown block/evade skills to neutralize it.

- Burst damage is way, way up in value. Putting healers into the game means you don’t wear anything down anymore, you either blow it up before their cooldowns come back up or you watch them reset.

- Excess cleave and reveal has turned Shadow Refuge from an essential skill for +1s and rotations into a death sentence, raising the value of all other out-of-combat stealth substantially.

That’s a meta shift that is very hostile to anything other than D/P, but quite favorable to D/P. Even is sword was operating at full power, it doesn’t offer the right tools to successfully counter the meta – it would have to be so powerful as to actually define the meta, which is very, very unlikely to happen.

Hesacon.8735:

That said, exotic berserker’s gear is dirt cheap from wvw vendors.

Amusingly, exotic Berserker’s and ascended Marauder’s armor have practically identical effective power – Marauder has a touch less with full raid buffs but a touch more with.

—

To the OP’s question, a bit of Marauder gear isn’t going to decrease DPS by a noticeable amount, and until you are very experienced with a raid the extra health will matter and will prevent being downed now and again.

Most of the talk of ‘gimping’ your damage is from people who have not done the math and just assume any gear that isn’t meta will be terrible.

babazhook.6805:

The duration of slow does not go up just because it hits more targets. Each taregt takes the one second hit.

So we’re clear, in the best case scenario (fighting two opponents), Dancing Dagger hits them both twice. Thus instead of hitting with a 0.6 coefficient and applying a 3s cripple, it effectively hits both with a 1.2 coefficient and applies 6s of cripple (to both targets).

Slow stacks duration, so in that ideal case (a 1v2, 2v2, or 3v2) Dancing Dagger would apply 2s (if 1s base duration) or 3s (if 1.5s base duration) of slow to both targets. You could spam that out to apply a lot of slow very quickly, since DD has a very short activation time and cheap initiative cost.

That’s one the big constraint on Dancing Dagger, and why it is so hard to balance – what can you do to it to make it more useful in against a single target, without making it go bonkers against two targets? It’s why poison (which stacks duration, so two stacks isn’t much better than one stack) and blind (where you only kind of care about stacking duration) are the conditions that make sense here – they’d be good vs one target but not scale up appreciably against two – and poison is a lot more thematic.

DeceiverX.8361:

No-cast-time IR is just way too strong, particularly with Daredevil, since it’d allow an Acro/Daredevil an effective stunbreak/disengage on average once every 3s

100% agree here. At most I think you can argue for evade frames on the activation of IR – that makes it still a proactive skill and not a get out of jail free card, but also removes the really demoralizing case of using it in response to a telegraphed stun and getting hit during the animation.

BlaqueFyre.5678:

reduce Larcenous Strike Initiative cost like it used to be

You think Larcenous Strike is too weak? I feel like it’s super strong when I actually get to use it. I’d personally look closer at Flanking Strike – that’s the weak point of the combo and not worth the initiative at all.

babazhook.6805:

I do not like the idea of Poison on dancing dagger simply because d/d already has poison on the AA.

S/D already has cripple on the AA! The main argument for poison (and not slow) is how it behaves if it hits two targets, especially what it does if it hits two targets with condition duration. Poison has a nice, solid effect if it only hits a single target, but if it bounces between two it doesn’t do a whole lot more (a little bit extra DoT). It’s nice for the healing reduction up front. Slow would be totally fine on a single target, but I think starts to go bonkers on multiple targets – even a 1s slow, if it’s bouncing between targets, turns quickly into 2s on 2 targets, spammable every 3s long term…add a bit of condition duration, and you’re looking at a skill that can maintain 100% slow uptime on a couple targets. With cripple on top, that’s looking pretty obnoxious.

babazhook.6805:

On The CND it my belief the INI costs are too high. Drop it by one and increase vuln stacks to 5 stacks. Add one second to base stealth. This latter to address the issues with the cooldown on attacks from stealth.

Yeah this is what a good start would look like. What makes CnD tricky is how intertwined its power is with the rest of the kit. For instance, the longer immobilize on Infiltrator’s Strike is a big boon for CnD on S/D, since it sets up your target to make CnD easy to land – that just ends up not being super attractive since S/D doesn’t have any tools to capitalize on it (none of the other skills involved hit hard).

Or, put another way, if the set-up / control tools are better, and the payload more attractive, CnD doesn’t have to be all that much better.

I feel like sticking poison on Dancing Dagger would be more than sufficient. Not every skill on your bar needs to be a monster – niche utility skills actually work really well on thieves since you can funnel initiative into your powerhouses most of the time, and the niche skill costs you nothing.

Cloak and Dagger is the one you want to be powerful. We’re long past the days in the first few months when it was a powerhouse because people couldn’t move or dodge worth a kitten . Stuff like baseline CiS would be nice, but wouldn’t really fix it; doing something like dramatically speeding up the animation would solve a lot of the consistency problems but could easily make it overbearing.

I feel like CnD doesn’t need a single big buff, but a bunch of small ones that together add up to a lot – drop the initiative cost by 1, baseline CiS, increase the stealth duration by a second – that make it a bit more reliable and a bit more powerful, without giving it a single dimension that becomes oppressive. That’s a lot harder to sell than a single, big number fix, but I think that’s the only way to eventually make it healthy yet powerful.

Bronze is serious business

tl;dr – market makers (“flippers”) increase the amount of gold sunk via TP fees. This, overall, makes gold more valuable and lowers TP prices.

HauntedCR.8350:

Why GW2 endgame gear is only about 5% better than the exotic?
That does disencourage people to work hard for it.
Such hard work should be rewarded with, lets say a 15% better item than the casual equipment (exotic).
Am I the only one that disagrees about lots of farming for only a 5% upgrade? Am I missing something?

You’re in luck! While the numbers on ascended equipment are only 5% higher than that on exotic, when you look at the impact of that 5%, it turns out that an ascended set of gear is 20% better than exotic equipment!

If the numbers on ascended gear were actually 15% larger than exotic pieces, then a full set of ascended would make you 70% stronger than someone in exotics – a power difference that is simply game-breaking.

I don’t think sword screams out with any one thing holding it back. It took some nerfs, and then has just been power creeped out over the past couple years.

The most obvious difference post-HoT is that your match-up against D/P thief has gone from a hard fight that was still largely skill based to one D/P simply dominates. Part of this has been the weakening of dodge spam (which is how S/D fought extended stealth), but the bigger issue is how S/D’s offense depended a lot on fishing with Infiltrator’s Strike and punishing when you hit the immobilize. Now that’s hard countered by Unhindered Combatant (or Debilitating Arc on staff)…meaning you really don’t have a threat to win with and you’re mostly just waiting until you make a mistake and die.

That to me means it needs power shifted elsewhere in the kit – 3 to 5 on S/D are marginal to weak, and only 4 is good on S/P (as you have no synergy with 5). Fixing that is a pretty big ordeal because it means turning those skills into the focus of their respective kits, which is not an easy thing that needs small number changes.

Platinum Doubloons are a pretty common product of throwing 4 T4 gemstones into the forge. It’s the best use of T4 gemstones AFAIK, and has reasonable profit per time.

You get about 6% health per 1% DPS you give up. It depends a bit upon build and game type (the buffs you expect to have). See below – you get about 6.2% health per 1% DPS in a PvP spec and minimal boons (just some might and fury), but only 5.8% in a raid build with full buffs.

Numbers are a bit better if you use a Ferocious Maintenance Oil (which you should in general – it’s about a 1% DPS boost in raids on top of the extra health) – you get a 6.7% HP boost per 1% DPS sacrificed using a PvP/solo build, and 6.2% with full raid buffs.

You rarely, if ever, want to go full Marauder – on the margin, each additional piece of Marauder is adding a lower percentage health increase (due to the higher base without it), while costing you more damage on the margin (due to a lower base). Plus the real value of additional health diminishes hard once you have enough to stop getting one shot. Still, the first few pieces are a pretty fantastic deal.

For the sake of comparison, you can get an additional 4500 HP or so (increasing your base HP from 11645 to over 16000) for the same DPS trade-off as taking Invigorating Precision.

The bend in the PvE chart comes from hitting the crit cap – there’s some leeway as I don’t assume 100% uptime on Keen Observer.

rennlc.7346:

Vault is really good against those who are not.

Staff thief is such an interesting case where the power distribution of the kit changes so radically with skill.

If you regularly play against bad players Vault is absolutely the major source of power for staff. Vaulting onto bad players wins fights as they sit there eating it helplessly – it does a lot of damage and easily wins against an opponent just spamming their buttons while you facemash 5.

If you regularly play against strong players, particularly stronger duelists, the keystone skill on your bar is Debilitating Arc. That’s what allows you to use Bound aggressively and frustrate less mobile foes with your slipperiness. Vault is still a valuable tool, but one that works more akin to a Backstab – something you +1 with, or otherwise use to punish a preoccupied foe that doesn’t have the awareness to track your animations and timings.

Complaining about Vault spam is like complaining about 5-trap Dragon Hunters. Not that you don’t have a point, it’s quite good at certain skill levels…but the complaint makes pretty explicit what your skill level is.

Astralporing.1957:

Remember, that by the time ascended gear got introduced (3 months after launch), there were already people with legendaries made. a year after launch i have personally knew more than one person that had several legendaries.
And getting the materials for legendaries now is easier than it was then.
Except for MC’s, that is.

Yeah. A big difference between now and then is the demographics of the player base – back when the game launched there were tons of relatively low attachment casual players cranking out mystic coins and dumping them on the economy. Now the population is a lot more mature and dominated by more high attachment, high playtime players.

Coins were a lot more abundant when there were a lot more casual players around that weren’t realistically in the market for a legendary to support each high commitment gamer working on their 6th.

Wanze.8410:

I dont think MC can be treated as a time-gated material for legendary crafting on an individual basis because you can buy them from the TP.

They aren’t time gated on a micro/individual basis. They’re time gated on a macro/population basis, which is important from a scarcity perspective. They ensure that the top end rewards can only enter the economy so fast and thus remain scarce and valuable.

That time gating gives them very important economic properties (they’re a negative beta good as far as the rest of the economy is concerned) which some of these recipes take advantage of.

They aren’t so much to slow down you from getting a legendary, but to slow down everyone else, on average, so you only see so many legendaries or other high end skins in the game and thus when you get yours it still feels awesome and exclusive and not that you’re just catching up with everyone else.

Wanze.8410:

And for legendary crafting, we should also consider the other faucets for mystic clovers (reward tracks, chest of legendary crafting materials) as a “source” of mystic coins.

Absolutely. The lowered demand for Mystic Coins (and the burden of stockpiling them for a legendary) is a big deal. It makes me a lot less concerned about the need to stockpile coins for a traditional legendary. However, the 2nd generation legendaries do demand a pretty onerous amount of Mystic Coins, and the price impact of that is at least within smelling distance of a range I’d consider unhealthy.

Basically, the higher the Mystic Coin cost, in the aggregate, the less time relatively people are having to spend farming / acquiring materials / otherwise playing the game, and the more they are having to spend just waiting for coins to be generated. It is a canary of sorts, if the I/O is balanced right.

mauried.5608:

what should their price be , and how would that price be determined, and how would the TP be regulated to prevent the MC price from rising above the regulated price?

Well the last part is easy – drop rates / sources can be manipulated to make them more or less common, which affects the equilibrium price.

As for what their price should be – Mystic Coins fill a particular and important role in the market. They’re a cornerstone alongside gold, karma, and laurels – gold is neither account bound nor time gated, while laurels are both account bound and time gated. Karma is not time gated but account bound, while coins are not account bound but are time gated.

They act as a brake if the faucets of farmable materials are too wide, or a signal of they are too stingy. As they are a time gated buy complementary good, by including coins in a recipe it’s putting an explicit time component into the recipe (so many player-days) and limits the rate they can be made. If mystic coins are worthless, then the set of time gated rewards is too narrow or coins are too abundant; if their price gets too high, it means either the time gates are too stingy or the material spigots are too generous.

As for what their price ‘should’ be? It’s a design parameter with a targeted value that’s a percentage of the total cost. If I was a designer I would aim for around 10% of the total cost of a high end item as mystic coins. If the price of coins is low enough that it’s below 2-3% (and thus an afterthought) I’m concerned that I’m giving out too many or that not enough recipes are using them; if the price goes above 20% I start to worry that too much of the cost is behind time gated materials and not enough behind playing the game.

To that end, Mystic Coins represent around 11-13% of the cost of one of the original legendaries, which I think is perfectly fine. They do however represent around 18% of the cost of an HoT legendary, which I think is pushing up into a range I’d be concerned about. That’s a price range where ‘waiting around for the time gated currency’ is starting to put the brakes on legendary creation, and that’s not a good set of incentives to have in the game.

So to that end I’d be concerned when the price of a Mystic Coin hits around a gold and a half, and my first impression would be to rip some of them out of the costs of HoT legendaries (or at least keep the requirements down on any new legendaries).

Justine.6351:

Some “lesser” black dyes look better because you can still see the textures of the armor. This new one probably falls into that category for some.

Yes. It’s a much higher contrast black than any of the other options – it’s lighter than black on ‘light’ parts of textures, while darker than abyss on others. Thus it dyes your armor black while also making the textures pop.

Narcissism and ignorance are a powerful, self-reinforcing combination – and yes, the internet has empowered it.

Naus the Gobbo.5172:

I didn’t even talk during the match.

Then congratulations, you beat up on someone so dumb they can’t even get their memes right.

Behellagh.1468:

That assumes you have plenty of mystic forge stones that cost you nothing and you don’t use up mystic salvage kits at an absurd rate.

Mystic Salvage Kits aren’t a bad way to burn your Forge Stones. You get around 40s per stone if you burn them that way, plus some additional convenience.

I rarely have many stones around as I use them to forge random exotics (low value ones plus account bound from fractals), which nets around 1g per stone (at no risk premium). It’s not an item with a ton of use though so I wouldn’t fault anyone for treating them as worthless.

Curiously, if you compare the Silver-Fed Salvage-o-Matic to the Mystic Salvage Kit at fair market value per stone it turns out you break even on it after 15k rares salvaged or so.

It honestly depends on what tier of play you’re in. In low tier games, people can’t stay alive for the life of them and high sustain, high damage team fighters carry the day. I’d much rather shamble mid on a warrior, guard, or necro and cleave everyone down with AoE because they’re bad, then win a simple, predictable match than try to create pressure with a thief.

It’s at higher tier games, where people don’t just die within 10 seconds of a fight starting, that classes like thief or mesmer start to shine – that’s where you need to create mismatches and push people out of position to win, which thief and mesmer excel at.

Plus, if you genuinely belong in bronze / silver the skills needed to yoyo and force bad match-ups on a thief or mesmer are going to be a lot harder to master, and less fruitful in the short run, than learning how to not die on a warrior or guardian and using one of those two sledgehammers to piledrive the trash on the other team. Low end players don’t have small windows of opportunity to capitalize on, they’re wide open doors of failure. Drive the biggest truck you can through it.

Black Scoutsman.5830:

I’m not saying the warrior isnt strong, but what exactly changed from the last 2-3 seasons on warrior that made them this way?

Big changes to amulets and support classes. 2-3 seasons ago supports ruled the day and the team with the most supports would just win teamfights. Also supports and team fighters would have tons of health and armor from tanky amulets, which combined with support let people absorb stuff like Warrior / DH / Thief burst and heal back up. Big team fighters like power warrior would also get shut down by spammed CC coming off of gyro explosions and random DH traps.

Now very few characters have any extra toughness (most run amulets with +560 vit and no toughness) and healing has been nerfed enough that you can’t just heal up the damage you take unreduced, so heavy hitters are killing people, and fast.

Same deal with the thief – a few patches ago it was pretty marginal since it couldn’t do anything in teamfights (no stability, got shut down by CC) and when rotating it would run into mesmer tanks that it couldn’t kill – they’d absorb the damage and heal back up before you could come back in again. Now they have toughness and have to spec into healing power to even try and heal back up, and can’t keep up. So what was a good matchup for the Mesmer a few patches ago, and a close matchup a patch ago, is now pretty favorable for the thief. Same with the revenant changes – beginning of last year revenant was nasty with tons of defense and sustain that thief couldn’t kill through, then ate nerfs that made the matchup favorable to thief if you could pick your windows right, to a gutting that makes revenant less of a threat and a pretty lopsided match-up for the thief.

So yeah, basically the defensive tools that stalled out and locked down the game got hammered, res control isn’t as crazy as it used to be, and classes that attack and kill stuff are way better now that we have an environment where people die.

Tapps.1479:

Regardless of this being true or not placement matches should not put someone on the top of the entire leader board. Surely you can agree with that?

The soft reset should probably include some amount of clamping to prevent issues like this. Still hard to solve. Seasons are really short given the volatility of a team game.

Honest John.4673:

You keep using the arithmetic average of your team’s mmr. ANet has stated they use the geometric average.

Do you have a quote for that? Doing a geometric average would be…weird, to say the least. It makes the assumptions of the matchmaker inconsistent (and assumes that skill differences matter less at high rating play, which is inconsistent with how things work in pretty much every other multiplayer game).

What is relevant, however, is the duo queue effective MMR boost; having a duo queue on your team means the algorithm thinks you are more likely to win. I can’t tell you exactly how much since the wiki is out of date, but it’s a pretty substantial difference.

Vagrant.7206:

ANet… I get the need for placement matches having value. But they have too much value. We’re mid-season now and this is still happening.

Remember that there was a soft reset, not a hard reset. We do not know the ratings those players had before the reset. If those are brand new accounts, sure, there’d be a problem if going 10-0 shot them to the top of the leaderboards. On the other hand, if those were 2400, 2600 MMR accounts before the soft reset, a 14-0 run to start the season would place them pretty highly even with pretty modest changes in rating during placements.

It probably just doesn’t update in real time. I bet it dumps the database to the ladder on set intervals to cut down on calls to a (pretty active) database.

Ryan.9387:

Can you show the math behind this? I’m curious how it is done.

Sure. Back of a napkin, first estimate the standard deviation of player’s ratings.

Crude way of doing it is to find X such that the probability of drawing a random number of 1749 or greater from a normal distribution with mean 1200 and SD X is 250 as likely as drawing a 2108 or greater from that distribution. I got a standard deviation of ~277.

pnorm(1749,1200,277,lower.tail = FALSE)/pnorm(2108,1200,277,lower.tail = FALSE) ~= 250

Then divide 250 by the probability of drawing 1749 or greater from ~N(1200,277).

250 / pnorm(1749,1200,277,lower.tail = FALSE)

Kreweless.2196:

Am I missing something?

You are missing that ratings are not uniformly distributed. They are normally distributed. The top 250 are the upper tail of that normal distribution.

Fitting a normal distribution with a mean of 1200 and the rank 1 and rank 250 ratings mentioned estimates a total ranked population of ~32,000.

OriOri.8724:

Wouldn’t it be better to just put an upper limit on how high you can place with your placement matches?

Don’t need an upper limit, but the volatility of the placements does seem to be too high.

steelheart.7386:

I don’t know about 50 but more then 10. Also, hard reset of MMR. There should be a reset of MMR before every season. They can keep the old MMR for matchmaking between seasons.

Hard resets are really bad ideas in team games – it takes too long for ratings to converge from scratch. A hard reset involves several months of bad matches, a few hundred games to sort out. You can endure that if you want to have year-long seasons, but even then, why go through a hard reset and force a couple hundred games of crap matches when you can do soft resets and get reasonable placements inside 50?

Smooth Penguin.5294:

• 1) What Anet thinks works doesn’t actually work. Meaning, allowing a game mode that’s supposed to be competitive shouldn’t reward losing.

Having the rewards at the bottom is actually really important. As the player base in general gets better over time, it’s the bottom that tends to suffer the most attrition, which makes it hard to matchmake new players who don’t know any better…so they get crushed.

It doesn’t hurt anyone to have a bunch of casuals kitten ing around in bronze. That’s an essential part of any game, a base against which new players can learn. And really, if you can’t beat up players with random builds just mashing buttons, why do you deserve to be any higher?

Smooth Penguin.5294:

• 2) Matchmaking algorithm still doesn’t work well. There should be no reason why 3 or 4 of the same class should be randomly spawned on the same team.

Matchmaking does try to break this up, but it prioritizes putting players of similar skill levels together over ensuring everyone is on different classes. Plus, population is not high, and people bandwagon onto popular classes. What do you expect it to do? You can de-prioritize rating even more to try and get diverse matchups, but once you have the DHs (currently) piled up in queue, eventually it’ll force those matches to happen.

The matchmaker has to work with what it’s got, and if you feed it a pile of DH’s it’s going to match a bunch of DH’s.

Smooth Penguin.5294:

• 3) Players are punished for things out of their control. You win as a team, and lose as a team. This motto would be true, if Ranked PvP were an actual team game. As it stands, being on a team with up to 4 random strangers of various skill level truly makes winning more of a coin flip than based on skill and strategy.

If you are playing at your level 7 or 8 out of 10 games are going to be determined by factors outside your control (other player’s skill/performance, class match-ups, etc). That’s still true even if you’re playing with organized teams – sometimes the key plays happen elsewhere. Sometimes it’s someone else’s job that is important. You are not always the difference maker. It’s possible to play perfectly and still lose.

But sometimes the important points of the match are on you. Sometimes you’re in the favorable match-up to press and win. Sometimes the opportunities are there for you and not someone else. Sometimes everything else is equal and your play is decisive. That’s when you need to step up and win.

When you are playing solo queue you cannot expect win 100% or 80% or 70% or even 60% of your games (unless you are one of the very best players in the game). Your goal should be to win 55% of your games, 5 out of 9. If you can win 5 out of 9 consistently you will rise in rating fairly rapidly. You can expect to win 55% if you are a consistently strong member of your team.

Smooth Penguin.5294:

• 4) No barriers to entry for Ranked.

See above, I don’t see what the problem of piling newbies into bronze is.

Smooth Penguin.5294:

• 5) PvP needs more reporting functions for AFKers and trolls. If someone is intentionally throwing Ranked matches, they shouldn’t be allowed to play anymore.

Agreed.

Smooth Penguin.5294:

• 6) High rated players treated differently than mid/low tier players. This is unintentional on Anet’s part, but their matchmaking system protects the higher rated players. They never have to worry about their teammates not understanding strategy, rotations, etc.

LOL, no. The player base isn’t deep enough to give those sorts of matches to top rated players. On top of that, there are a lot of very talented one tricks out there and pretty far up the ladder. There are some really masterful players that hang out in dueling rooms, that are fantastic duelists, that know the ins and outs of those fights and just dominate their 1v1s when they play ranked – but when they get collapsed on or have to teamfight they seriously have no idea what they are doing. There’s a whole godkitten ecosystem of that in the higher tiers.

Plus, and I don’t know how this is lost on so many people – top players often make smurf accounts for the sole purpose of playing in bronze against new players. They don’t ‘get stuck’ there because their teammates are bad. They run their opponents in circles and carry super hard. I am by no means a top player, and back when I played frequently I’d jump into low rating games and just farm people. It doesn’t matter how bad your teammates are when your opponents are down or waiting to respawn. You gotta farm harder than your teammates can get farmed.

Astralporing.1957:

Sure, but then you’d need to ask yourself a question – why would they start to drain down? It’s not like there will be more incentive to selling.

They’d drain down when the rate of consumption exceeds the rate of generation. Measuring the volume of trade doesn’t matter when it comes to inventories – moving them from one person’s bank to another doesn’t change the total number sitting in all banks.

If inventories are way down because people have been consuming a lot of old coins by making skins in the forge, and that scarcity is driving prices up, then there may be reason to intervene. On the other hand, if prices are up because people are simply demanding higher prices, while building up inventories, there is no reason to intervene.

If I were them I wouldn’t be targeting prices as much as watching inventories; until those start to drain down there’s no need to take action.

I’d also be concerned about how much of the price surge is being driven by reward creep in the rest of the game. As the rest of the components for legendaries / high end skins become more common there’s more pressure on the mystic coin supply. If this is just the initial burst of coins from the start of the game being drained down, that’s great for the market long term, but if there really is a glut of other materials and coins are poised to become a gating item then something should be done.

Shaogin.2679:

wondering how your teammates could possibly be this much worse than the enemy team.

…they aren’t. That’s a really small gap, especially considering that everyone on your team just dropped 10-15 rating and the opposing team just gained 10-15 rating. We’re complaining about team MMR differences of 20 points and within team MMR differences of 100 points now? 100 points is, like, really close to the standard error on your MMR in a 5v5 team game.

Polismassa.6740:

2) Your rating is not a pure measure of your skill MMR in a team based game is, and always will be a measure of your ability to win matches.

Just to build upon this:

There are a lot of different facets to ‘ability to win matches’. Players can be good at winning duels or winning team fights or picking winning fights or fighting on the right points or rotating between points or playing good builds for the meta or swapping to the right build for the match-up or filling in effectively to balance your team composition or playing hard every game or not tilting or encouraging teammates and keeping people focused.

Or a whole lot of things I didn’t mention in that run-on sentence above.

The difference between a pro-level player and an experienced but middling journeyman (say, gold level) player isn’t that the pro is a bit better at all of these things. It’s almost always the case that the pro is good at almost all of these things, while your middling, experienced player is good (sometimes even pro-level!) at some of these things, and just a dumpster fire at others.

Good but not great players include people that can beat everyone they face in a 1v1 but are constantly engaged in bad fights (trying to kill a guardian on a fully capped point with their thief), people who can’t win a duel for the life of them but pick teamfights they can contribute to well, people who know a single trick (condi mesmer that always goes far!) that sometimes dominates and sometimes feeds, people who often play well but tilt and throw as soon as something goes badly (thus losing games that could be recovered and won).

The really insidious part of this? If you are one of those players, you’re probably really focused on the parts of your game that you are good at, and blind to those that are holding you back. You’re also blind to the good things your teammates are doing (that you are not good at), while their flaws (in areas that are your strengths) are on full display.

Which is why you think your teammate (with the same MMR and similar match win performance) is terrible, and why that teammate thinks that you are terrible. It’s why there’s a pretty big range of play (generally the 75th to the 95th percentile, roughly) that is just an utterly miserable rage-fast, a bunch of skilled but fundamentally flawed players picking at each other’s faults.

If you want to get better, don’t look at that ‘scrub’ with the same rating as you in disbelief. Figure out what he’s doing that you aren’t – because, odds are, that’s what’s holding you back.

Sky.7610:

This will drain some gold from the richer because the TP fees cant keep up inflation.

Sure they can. In fact they keep up with inflation by construction.

Illconceived Was Na.9781:

No, it’s a terrible indicator. The gold:gem ratio has frequently gone up or down independently of the money supply.

In fact, it’s not an indicator at all. Gold:gem ratio is as much a measure of gold inflation as it is a measure of US dollar inflation or Euro inflation. The two are not completely unrelated, but the correlation is close to zero.

Gerrand.3085:

Evil idea which would force most people to diversify their farming habits – to ensure their wealth isn’t locked into an single currency they could be robbed of.

An idea which makes your ‘currencies’ no longer currencies. If you put something like that in the game people would store their wealth in ectos or powerful bloods or whatever the game does not consider a currency, as it would then be more of a currency than gold. You know, the situation in a huge number of MMOs with failed primary currencies.

It would also make the trading post largely unworkable, since holding gold is a liability…which would make just about everyone a whole lot poorer.

This sort of thing benefits long term, dedicated traders that understand the systems, but hurts the casual player that doesn’t have time to dig into the nuances of trade and just expect gold to work.

I like staff for solo queue. It’s a flexible build that can fill in gaps / cover for big mistakes by your team better than D/P – it’s much better at holding/stalling caps in a duel and has less one-sided match-ups, so you have more potential to outplay a condi mesmer or a guardian than you would on D/P. When the skill differential on your team can be all over the place and the team comps have gaping holes I find staff to be valuable.

In competitive 5v5 team play? Forget about it. You should never be holding a point 1v1 waiting for people to rotate to you – if you are your teammates should be fired. +1s matter more than dueling power when you can count on your team to contest multiple points at once. Stealth jukes matter more when everyone is watching and communicating what’s going on. D/P is very clearly better when you get to focus on what you do best, rotating and +1ing fights your teammates are already engaged in.

But when it comes time to carry a bunch of randoms eating paste and dogpiling onto mid, I’d much rather have staff for its point control and 1v1 power.

Nimrud.5642:

1. Is 100% crit chance the sweetspot for all power-based builds and classes?

In full raid buffs, yes – your power and ferocity are boosted enough that precision is the most valuable offensive stat when in that state, up to the 100% crit chance cap.

Thus it is advisable to build power-focused raid characters to hit 100% crit chance, without going over.

Without full raid buffs power will typically out-perform precision.

Nimrud.5642:

- staff ele, thief and radiance guard are on 98.x% – using the suggested build and gear

There is a display bug on the character panel; your actual critical chance is 1% higher than what is displayed. Hence these are all 99.x% builds – and you can make up the difference with a couple precise infusions, which are not listed.

Nimrud.5642:

- honor guard, glyph power ranger are on 83.x% and 92.x% – using the suggested build and gear

Honor Guardian uses an Assassin amulet, Accuracy sigil, and a Maintenance Oil to hit 100% crit chance.

Power Ranger just uses the Accuracy sigil to reach it.

All their listed builds are already optimized to hit 100% critical chance.