Glorious Alpha Two Testers!
Phase I of Alpha Two testing will occur on weekends. Each weekend is scheduled to start on Fridays at 10 AM PT and end on Sundays at 10 PM PT. Find out more here.
Check out Alpha Two Announcements here to see the latest Alpha Two news and update notes.
Our quickest Alpha Two updates are in Discord. Testers with Alpha Two access can chat in Alpha Two channels by connecting your Discord and Intrepid accounts here.
Phase I of Alpha Two testing will occur on weekends. Each weekend is scheduled to start on Fridays at 10 AM PT and end on Sundays at 10 PM PT. Find out more here.
Check out Alpha Two Announcements here to see the latest Alpha Two news and update notes.
Our quickest Alpha Two updates are in Discord. Testers with Alpha Two access can chat in Alpha Two channels by connecting your Discord and Intrepid accounts here.
Comments
I agree that balance in a game like that requires that no functional group have less than 6.5:3.5 matchup anywhere in the tierlist.
You can assume that when talking to me specifically, I am talking this level, same as you.
I would also accept either position on this type of situation for a premise. That a game that contains this character is 'unbalanced' because the number of matchups they have where they are considered good is too low, or that the game is balanced because none of the matchups are considered low enough to be unwinnable.
And this is in the game where all the stages are the same, the starting conditions are the same, and player skill, on average, doesn't cripple the execution of most characters.
I don't think anyone should come to the conclusion that they should play whatever comp they want. I think they are concluding that the comps that the designers have told us are acceptable, will be within the ranges they are okay with.
I expect that Argent, Necromancer, Minstrel, Warlock, Soulbow, Assassin, Dreadnought, High Priest (Tank/Bard, Summoner/Cleric, Mage/Summoner, Ranger/Cleric, Rogue/Rogue, Fighter/Tank, Cleric/Cleric) will be the 'best comp', in any corridor heavy dungeon, on the order of 'having few if any bad matchups', and having great matchups against groups without Fighters, doubly so against groups without Mages, enough to make it more lopsided than the linked data.
I expect that Paladin, Oracle, Siren, Highsword, Spellsword, Bladecaller, Duelist, Keeper (Tank/Cleric, Cleric/Mage, Bard/Tank, Fighter/Cleric, Fighter/Mage, Fighter/Summoner, Rogue/Fighter, Tank/Summoner) will be stronger against that comp above, in enclosed spaces where they normally dominate.
I expect that group A will change the matchup if they realize things are happening by doing their best to fight while rapidly retreating to a larger room with a side pillar or column to lessen or negate the advantage of Group B's composition. I expect that sometimes, this will be the wrong thing to do because Group B will have spec'd individual characters to have unexpected strengths for their class, making the attempt at countering useless.
It is within these expectations that we find the things I'm ... 'challenging', your perspective with?
Matchups change by location.
Matchups change by skills taken.
Matchups change by cohesion and tactics.
Smash is actually one of the least flexible games in this regard, so I think there is a small possibility that your points are being disregarded even within people who know what you mean, because it's a case of 'why would it be that bad though?'
More popular games are often less balanced because the standard distribution favors games where the answer to 'what's the best X?' is easy, when selecting for popularity.
So I need to know 'where to meet you', let's be 'antagonistic' for a while in shorter posts and see where we get.
Basically we don't know. I have heard things like people may be insta-killed or teleported out if they enter the area without a defender or attacker tag. For all intents and purposes this would be like an instance really, but people would be able to stand on the sidelines as spectators and look in.
I am not even sure if node sieges are going to be capped at 250v250. A metropolis might have well over a thousand citizens for all we know, and only letting a quarter or less defend their home seems wrong.
As for the balancing part, I am pretty sure they've said that balance is more focused on 8v8 in some interview or dev stream.
It could be done in a manner similar to phasing, with each phase it's own battleground with it's own player count and it's own objectives.
Or it could be that the area of the node is broken up in to perhaps 4 areas, each being it's own battleground with it's own player count.
Each of these could easily see a single siege have 2,000 players per side while the cap per battleground is 500vs500 (which is the goal).
Or they could do both concurrently. Or neither.
Who knows...
Castle and node sieges have declaration mechanics. wich is tied to a signup. So i assume its limited by signup.
250 vs 260 might seem like a lot, but remember there also working on 500 vs 500. Now at the start that would most likely be enough. But depending on server and node popularity it might not.
2 examples: A streamer with a large group of followers desides to play and on that server they control a node. a group attacks it and everyone of its followers wants to attend. the 250 cap on the defending side is easily reached. In this case it could be detrimental. the cohesion is far lower, and not everyone is max level, mos can't communicate with eachother etc.
The next day the same group decides to retaliate. they all log on and theres thousands of them. the zerg quantity would be able to overtake any node. but only 250 of them can attent the siege, the sieged node having far few numbers can actually defend.
Note artificial limitations on numbers are there for one of 2 reasons.
1) to ensure that the game can handle the in and output. so the game remains playable for all players
2) to ensure that pve encounters can be finetuned..
That may change - I am the first to say that nothing is set in stone. However, the current thinking from Intrepid is that sieges are open world.
They have a few plans for how to restrict non-registered players, but the current intention is open world.,
If it helps, if someone were to create a matchup chart for melee on specifically Final Destination it would look different (and more marth favored) than the one we have. Matchups definitely change by stage (which is like location in an MMO). But, a best of 5 set starts by striking down the stage list, playing that out, and then the loser counter picks stages. It's a little more complicated than that, but that's the gist.
Matchups for sure change by skills taken (since your character in an MMO maps to your build, which is your class + skills + weapon + armor + augments + etc). Need to have a build, which includes skills, before you can think about a matchup, generally. I don't think it makes a whole lot of sense to say "what's the templar vs paladin matchup?" when there are hundreds of thousands of templar builds and hundreds of thousands of paladin builds. That would be an incredibly ambiguous question!
But! We wouldn't traditionally say that matchups change by cohesion and tactics. I believe this is akin to saying that matchups in a fighting game change by how the players approach them. Traditionally, the figures cited in matchup charts represent (as far as I'm aware, and unless specified otherwise) "top players playing their characters to the best of their ability using the most effective strategies and tactics".
You could intentionally use garbage tactics and then you'd have a garbage matchup, and we could create a matchup chart to reflect how you have garbage matchups when you have garbage tactics, but I don't think this is useful. Instead, these charts are supposed to reflect the state of the art.
That said, as I mentioned to ptitoine a little earlier in the thread, some folks find it useful to create separate charts for "intermediate players playing their characters to the best of their ability using the most effective strategies and tactics" and "beginner players playing their characters to the best of their ability using the most effective strategies and tactics". These sorts of charts can be really useful for answering questions like "If I'm a beginner, and my friend is intermediate, and I'm willing to put in enough effort to get better than them, what are the good characters to do that with?"
I totally understand the proposition for the group A / group B example. We're saying that there are some stages where group A has a favorable matchup, and some where group B has a favorable matchup. That completely tracks.
I'm saying that there's a few pieces of trouble. The first (assuming that we can somehow move past this aside with siege open-world-but-not fuzziness), the siege of a particular node is a single "stage". That stage is huge but it's one stage. It might have multiple areas that where group A is stronger than group B, but at the end of the day, one 250-man team fights another 250-man team and the node is either sieged or it isn't.
It may turn out that a particular bridge is the crux to the whole thing, and so you need to optimize you whole setup around being able to attack/hold that bridge. It may turn out that there are multiple ways to play it so you need to be well rounded. It may turn out that there are tons of different objectives you need to split up into a bunch of sub-teams. In any case, you still have to choose a 250-man team (which maps to 1 character), and they have to do the same, and then you play 1 siege, and 1 team wins.
The stage for node-43 will be different than the stage for node-17, so it makes sense that you'd make a different matchup chart for that. I also wouldn't be surprised if the numerical entries were extremely similar.
Put it this way: say that the devs lifted the siege part of the game out into it's own lobby game and let players queue for that all day. There are no grinding elements, so you're free to pick whatever skills, class, gear, etc you want for your character. Like playing infinite siege on a public test ream, except it's always for node-43. Each time a team wins a match, update their matchmaking rating. Don't update balance for 5 years. At the end of 5 years, give the top MMR team 70 million dollars, give the 2nd place team 30 million dollars, and the 3rd place team 10 mil.
Do you predict that a meta will form and that the top teams will gravitate toward the same small handful compositions, or do you think that folks will be playing all kinds of stuff because everything is "balanced" and it's all up to preference/taste?
Relevantly, there will be arenas.
It's looking like the brackets will be 1v1, 3v3, 5v5, and potentially 20v20, with MMR. This is something I'm super familiar with. I went to MLG Orlando for WoW in 2009!
Notably, 8v8 isn't in there, but 3v3 will be (and that's the bracket I'm most familiar with). 3v3 is explicitly not being balanced for.
Example, Japan has a format for certain games where they play First to 1 instead of first to 2 and the meta is different and the tierlist is different because 'when and how you guess against your opponent's psychology' is crucial. Reasoning given is that First to 1 involves more 'soft skills' (intimidation, psychology) and luck (the optimal strategies for most games in the long term are defensive so that you don't tip your hand by doing risky things your opponent can learn and capitalize on later).
This happens in Paladins all the time. You go for a point-control-win in a way that is only effective because your opponent has never seen your strategy for doing it (this is possible by the aforementioned, always lose the point, always defend the opposing push, until the score is 3-3 and you HAVE to win the point-control, but your opponent has never seen your strategy for actually successfully taking the point.
If you take everything Intrepid is saying at face value, I'm saying it's possible. And my position here wasn't to tell you 'this will be balanced', only that it 'can be balanced', and that the examples being used, may not be resonating with those you're trying to talk to.
SFV, by most standards, has no matchups that even reach 6:4. Pros make decisions on things like 'Gill is better against Akuma than Urien, but I feel more comfortable with Urien so if he picks Akuma I will probably still play Urien'.
I believe the Arena will have a meta. SFV has a meta. That meta is different by region. In some parts of the world I choose Chun-Li for tournaments because opponents play footsies. In others I choose Kolin because opponents play defensively. My opponents could choose the exact same characters in both regions and I would, while going meta, explicitly be better off with different characters even if they attempted to change to counter which one I played.
Because you can't counterpick me by switching to a 'bad matchup' unless you can also naturally play in the relevant style to make that matchup bad enough for me, and it would take months of practice and personal conditioning to do so. It's actually neurologically near-impossible. The mental 'space' required to play against my Chun and my Kolin are practically exclusive, to the point where I could 'start as one, see if I can win, push a person into that mindset, and then switch to the other and win before they could change back'. This is another common pro tactic.
So, I'll engage some disingenuous hyperbole. You're an internet person enough to know the 'style' this is normally said in.
"It's because you play Smash and other unbalanced games that you're used to assuming that it has to be that way. Go play some Paladins or something."
At that level of game complexity... I believe that players will play whatever they personally are best at, if the game is balanced enough.
You have to 'define' what amount of 'all kinds of stuff' counts as balance to you.
To me, if 4/8 classes within most or all Archetypes are acceptable within an 8v8 team within a Node Siege, in over 100 combinations, this is balance. Imbalance comes generally in two places that are relatively definable, and the core systems of Ashes do not, in themselves, imply the existence of either. You're familiar with both, as a Smash player.
1. Group A enters situations in which Group B's advantage is practically assured and this situation is an unavoidably frequent part of the gameplay, or not bypassable due to situation/encounter design.
2. Group C is just a flatly better version of group D in any part of the gameplay that can be engineered easily by the opposing group or has a very difficult counter when used against an opposing group, i.e. their 'alternative strengths', if any, never matter.
Of the 51 playable Champions in Paladins, the above is true for 10-12, generously biasing toward 'absolute meta, no personal style synergy'.
If, of the 64 playable Classes in Ashes, 56 are balanced, or of the 64^8^8 (classes by probable number of builds people would consider viable, then restricting to only comps that contain one of each Archetype), 3.2e+115 of the 3.940201e+115 are balanced by extension of their design, is this good enough?
One question about your concerns for balance:
With diverse teams, don't the stats tend toward gaussian distribution as you increase the size of the team?
As for sieges, I also expect the tactics of siegefare to develop more complex strategy, e.g.
- You find that your team has an excessive number of mages, so you assign a batallion wing of mages to teleport-flank the enemy to burst down a key objective at a critical time.
- You find that your team has too many tanks and not enough healers, so you plan for the extra tanks to disperse and force small skirmishes away from the efforts of the main team so that the enemy has to thin their lines equally.
Is this a form of combat balance that you can see happening? I think this is also why the non-class-specific siege weapons will help to curb inbalances, because non-essential classes can jump on the ballistae/trebuchets/etc. (Presumably there are going to be more of these types of activities)I think it IS inevitable that certain builds of the same class/archetype are going to perform better in sieges, against others in bossing, etc. due to the nature of those activities being so different - ESPECIALLY if Ashes is opting for ye olde hate/threat system.
All this, too. In a game with weaker matchup balance, one does not have to consider the 'cost of adaptation' which I alluded to above. In games where matchups approach even, then complexity increases the cost of adaptation to the point where 'choosing to play a non-meta strategy', even if it is not explicitly a counter, still forces the matchup to 'even'. Your opponent is paying the cost of adaptation and you are not.
I think you know this too.
Games these days are balanced better. Some aren't, and those are more popular in general.
Ashes apparently doesn't aim to be 'popular' first, so maybe it will work out.
This all tracks to me - I think it makes sense to potentially create different matchups for best-of-1s than for best-of-50s (gimmicky characters excel at bo1s whereas solid characters excel at bo50s, at a high level).[/quote]
This is where I'm getting confused. If you're agreeing with the definition (no matchups worse than 65:35), and you're understanding the size of the matchup chart in 250 v 250 sieges (quadrillions of matchups), the claim is that there's a chance that none, (even after we apply reasonable filters like "has to have a reasonable mixture of tanks, dps, and healers"), will be worse than 65:35, even the "worst" of those vs the "best"? The probability is technically nonzero! I just think it's incredibly low.
That's a fair point. Balanced games can have a meta, and imbalanced games don't necessarily have to have one. Just depends on how much the community cares about playing optimally and whatnot. SFV's ability to have such an even matchup spread with that many character unique characters is a game design marvel.
Hah! so in classic internet rebuttal: "Did you miss the part where I was a masters sc2 player, and grand champ rocket league player?"
On a more intellectual level, I'm not assuming that it has to do this way based on narrow life experiences derived from playing melee. I'm looking at the number of matchups that Intrepid will have to deal with in the 250v250 bracket, how they plan on controlling those matchups (they don't control players builds), how they plan on balancing those matchups, and estimating the probability that they all come up relatively even. Estimations are vanishingly low.
This pre-supposes sufficient balance though! Under that pre-supposition, I totally agree. It's that assumption that I have trouble with.
This depends on what sort of question you're trying to answer!
If your claim is "you can pick any build and find success with it", then nope, not good enough. Using the first set of numbers (because they're easier to type), you still have 8 "trap" classes. This is still way better than a lot of games, but that's still 8 trap options nonetheless.
If you're willing to look into the shades of gray instead of the marketing bullshit, then the outlook gets a lot better. Instead of "balanced or not balanced" you have "better balance" and "worse balance", then this is great!
I think having a game where 87.5% of the options being balanced would be incredible! Or a similar stat for total build count! I also (perhaps pessimistically) think that number will be a lot lower, though my standards are different than most folks.
If your team is very diverse, then everything basically averages out. If it turns out that your 250-man team has ~4 of each class, and each of those people are playing a different top-tier-for-sieges build, then that would be pretty wild and a sight to behold.
On the other hand, here's what I actually expect to happen (along these lines). Folks figure out you definitely need all 8 archetypes. In the particular balance patch, there's a particular build for ranger/mage that makes it the most effective single target ranged damage dealer (by a good 15%) with no appreciable downside, so you stack a bunch of those. In fact, there's no real reason to take any other type of ranger.
It turns out that in Siege settings, fighters get melted by the focus fire of the aforementioned ranger/mages (who can enchant their arrows with magic to help do armor pen), so the fighters that can survive the best are fighter/tanks, and since in this balance patch, the survivability increase is around 15%, and the dps loss is only around 4% relative to fighter/fighter, you stack a bunch of those for that slot.
For you mages, due to an exploitable bug involving an advanced technique the community is calling "song-cancelling", you want nothing but mage/bard, etc.
So, you're have your full 250-man team and it represents all 8 archetypes, but it only has maybe 15 different classes and 20 different total builds, because those are all the strongest in the current balance patch. If you can get this full team together, the build strength alone of these characters is absurd. I've seen it happen in ranked battlegrounds in WoW when boomkins were strong and it's oppressive. Enemies just melt and your team feels invincible.
Then, a couple of things may happen.
Either Intrepid either nerfs it, or the players straight copy the strat, or they find some way to make an even better comp. If they picked the third option, the cycle repeats.
Note: this presupposes the existence of builds that are stronger than other builds (which I don't think is a wild assumption)
Here's this principle playing out in a PvE setting. Note in the bottom left the number of brown nameplates. Those are all "fury warriors", which is the highest dps dealing build in the game.
It wouldn't be quite so bad in PvP - Molten Core has no way to "counter" a bunch of fury warriors. But the idea is the same - you figure out whichever builds excel at the format the best and you play that. If your opponent somehow figures out how to stop you, you figure out whichever builds excel best at not getting stopped by what stopped you and you play that. Eventually you either hit some sort of rock-paper-scissors situation or you find the set of teams that all go relatively even against each other.
You hope that set of teams is large and includes many classes/builds, like SF5 does, and not small like smash melee or WoW Classic.
I think this conflates what a matchup chart is. Matchup charts don't talk about particular upcoming matches between particular players who may be unfamiliar with the matchups. They represent (again, as far as I'm aware), "top players playing the matchup to the best of their ability using the most effective tactics available".
Not "particular players playing unfamiliar matchups having to adapt to off-meta strategies"
So, while fox-kirby might be 75:25 because once everyone involved is super good at the game, kirby doesn't have a lot of great options and the gimmicks stop working, that doesn't mean a fox unfamiliar at the matchup will beat a particular kirby who does nothing but play his brother the fox main all day.
And here's where we are trying to reach a consensus, so again, being antagonistic because it makes this a bit faster.
SFV is not a game design marvel. SFV is the result of just proper design which I feel like I could teach anyone with a strong enough interest in design how to do within a week. Changes to SFV for balance are predictable among top players to the point where some of them just 'prepare' or 'get to complain when Capcom takes six months to make the same changes that the community as a whole, foresaw'.
I'm not saying Ashes will be balanced. I'm saying that you, or I, alone, could balance it to the point of 65% easily, 87.5% probably (or together).
I still disagree highly about the amount you're applying to cost of adaptability in well balanced games.
This is a match with a new and unfamiliar character, but it's not just a new and unfamiliar character, it's the specific way that Momochi chooses to play this character. It's not that the way Momochi is playing is a particularly optimal method for this matchup compared to other methods, either. It's that the skill type and the player's style matter so much more than the methodology that Daigo could not deal with this in the moment.
Over 6 years, thousands of games, months of training, sure. But when the game is complex enough,, there are simply too MANY of these for anything beyond a meta-stable meta to form.
Old games are not like this.
Almost all recent Fighting games are balanced enough that 70% of the cast is outright tournament viable, though some are obviously stronger than others.
Smash is the notable exception to this. So yes, I'm still 'calling your experiences into question' here. If you're worried that they will fail, or that they will 'claim everything will be perfectly balanced when that's impossible', then I concede entirely. But I'll oppose anything to the effect of ''there are too many to do it', because that's false.
I played a ton of SF4 (placed in Atlanta locals, traveled to majors, netplay warrior, etc). I haven't touched 5. I have also played a bunch of Strive. Say that Ryu was struggling in SF4 with the Akuma matchup because Akuma could punish c.mk -> fireball on block, on reaction with super (this wasn't the case, but say that's an example). Then, capcom goes as makes it so either Ryu's blockstun on fireball is a few more frames so that ryu can jump out after the super flash, or they change it so that Akuma's super flash has more startup or travels slowly enough so that ryu can jump out (or any number of other things to make that matchup more even).
This ripples!
If the rest of Ryu's matchups were even, and Ryu got a buff to his block stun on fireball, now he potentially has winning matchups. If the rest of Akuma's matchups were winning and he got a nerf to his super, now he has potentially losing matchups. This gets harder the more matchups there are.
As for your other points, I'm curious as to how many teams you think there are (ballpark) in the 250v250 bracket (which map to individual characters in sf5), and how many matchups that creates in the 250v250 bracket. Then, if you make a balance change to one build (which a member of a team), how many matchups that one balance change influences.
To be clear, I think this can be done, I just think it would need to be done by a machine and not a person. I would love to see extremely frequent balance updates (daily, potentially), that are small tweaks based on automated metrics fed into some sort of neural net. I would love to see balance proposals created by such a neural net, curated by developers and then proposed to a council of top community members that have been voted in by the community at large to represent their interests. Then, this council (like the US senate) votes on these changes and the game progresses. I wrote about that here.
I agree - this is all @beaushinkle 's fault!
From the way SS still talks about it, it doesn't seem like they've finalized the mechanic, just yet. A city can have more than 250 citizens - but it would be a tad disappointing if the mechanic was to queue and then *be chosen* from the highest-ranking city official online.
Presently, yours truly suspects that it'll just be first-come-first-serve from the citizens up until so many minutes before the siege is to begin - and then anyone who's not a citizen can join in to help.
Heh - or to "help", as the point of your thread would infer.
This can be done by people. NNs help. You can surely infer why I in particular say this.
Quick napkin math. In your 250-man group, say that you want 15-35 tanks, 15-35 healers, and the rest as dps.
There are 8 different tank classes, but how many different viable tank "builds"? 40? There are 8 different healer classes, but how many different viable healer builds? 40? There are 48 dps classes, but how many different viable dps builds, (to keep it in conservative numbers), 240?
If we were trying to make a 3v3 team consisting of 1 tank 1 healer 1 dps, we're already up to 40*40*240 = 384k different teams.
If we were to make a matchup chart between just those 3v3 teams, that matchup chart would contain (384k ^2 - 384k) / 2 = 74 billion matchups. That's just 3v3. Am I misunderstanding something, or are you suggesting that a human being can balance 74 billion matchups by hand? If you're thinking about the problem in a different way than I am, can you elaborate?
If you go from 3v3 to 4v4, the numbers explode (1 tank 1 healer 2 dps is 92 mil teams and 4E15 matchups). 250v250 would overflow my calculator.
It is because balance is a fractal not a jigsaw. The NNs help finding fractal breaks, but one can visualize the fractal itself.
The specific class make up of a siege is going to have less of an impact on who wins than the ability and effectiveness of the leader on each side.
A siege isn't going to be a mass of people attacking a mass of other people, it is going to be a mass of people being broken down in to smaller groups to each perform a specific task (sieges have objectives).
A matchup chart only works when you have two groups of players with the singular goal of defeating each other. When you factor in other goals, a matchup chart simply isn't appropriate to use.
Even if we assume 250v250, and even if we assume this is closed, if I am a better leader than you, and I send the appropriate number of players to each objective as it comes along, forcing you to have to redirect your forces (which means I will have 250 people in play at any given moment, while you may have 220 in play and 30 in transit), then I am going to win. Class ratio be damned.
I get that you are trying to use tools that you are familiar with from games you have played in the past, but Ashes is not like the games you have played in the past, and so can't be explained using the same tools.
I really don't think you understand matchup charts
What is it you think a matchup chart is able to tell us in large scale PvP?
Or even in 1v1, for that matter.
I don't think there is anything useful to be gained from a 1v1 matchup chart in a game like Ashes.
As an example, lets assume that we have a comparison of fighters and tanks. We pulled this data from the server, as there is no other way to get it, and the data only includes times when these two classes meet up with full health, full mana, and no other entities deal any damage to either party. All of these things need to be true in order for a data point to be able to be entered in to a match up chart.
Lets assume that fighters win 70% of the time here.
Now, common wisdom would suggest that this means fighters are more powerful than tanks, since they win more. But then you need to ask yourself - what was each player doing?
If the tank was temporarily on their own, perhaps they were running a spec that is better suited to being in a group. If this is the cased, then almost any other class would be able to beat them.
That doesn't mean tanks are less powerful, it means the build that tanks in groups use is less powerful if they are not in a group.
Basically, this matchup says nothing at all about any other builds that are available to each class, and if you are going to assume a simple matchup of a class vs another class are actually valid, you need to assume an even spread of spec's is present in your data.
That is an assumption that you simply can not make.
As such, you can not assume a matchup chart is valid in terms of one class vs another in Ashes.
The same can be said of gear.
When you start getting to large scale PvP, the leadership capability and structure are more important than the class makeup of each side.
As such, comparing the class makeup of each side is not only impossible unless the leadership of each side doesn't change, but is actually pointless even if possible, as you are trying to make a comparison of the second factor rather than the first.
That is a matchup chart.
What I talked about fills in two data points on that chart. Fighter/tank is 70/30.
Not sure what you're trying to get at with this post. I was talking about a single (well, two)data point in a chart, and you are trying to tell me that what I am talking about is not a matchup chart. If that single (double) data point is not able to tell us anything in Ashes, then the chart as a whole isn't able to tell us anything.
Matchup charts are theoretical. As I've mentioned several times throughout this thread, each cell answers the question "if a top player playing row_character using the best tactics and strategies available to them plays against a top player playing row_character using the best tactics and strategies available to them, then we expect the player playing row_character to win cell_value % of the time"
If a new strategy is discovered, the matchup chart is updated before the actual matches between top players even happen.
It is an understanding of the state-of-the-art.
To understand how this applies to groups of players playing a team, you would get a statement like this:
"if a group of top players playing row_team_composition using the best tactics and strategies available to them plays against a group of top players using column_team_composition using the best tactics and strategies available to them, then we expect the players playing row_team_composition to win cell_value % of the time".
In a game like ashes, we can inform our understanding of the most effective strategies and tactics, and inform our analysis of a particular matchup like "such-and-such tank build vs so-and-so mage build" by looking at a lot of VODs of tank vs mage fights.
Maybe we watch hours of footage and we notice "oh but the tank could have done thing thing, and they didn't, and that's why they lost, but they could have won here", and at the end we conclude that the "such-and-such tank build vs so-and-so mage build" matchup on an open field is 35:65 in the mage's favor.
This 35:65 conclusion might be pretty different from the historical statistical sampling, but the historical statistical sampling also does not at all represent "top players on both sides playing as well as possible".
Does all of that track?
Explain to me how you would theorize a matchup chart in Ashes. Would " the best tactics and strategies available to them" be based on the idea that they know who they are going up against? Because that drastically changes things.
But then what is that going to tell you about anything actually real? If my matchup chart against one class requires one build, and against another class requires a totally different build, which build do I take with me on a siege?
I can't take both.
How does the matchup chart account for this?
I mean, the chart is all about " the best tactics and strategies available to them" in a theoretical sense, and that will always fall totally flat in a practical sense.
I mean, you are creating a matchup chart to inform your choices in regards to builds (individual player or siege, what ever). How can that work if the match up chat uses " the best tactics and strategies available to them"? If you don't know what it is you are going to face when you leave town, how can you know what " the best tactics and strategies available to them" even is, since it is dependent on your target, and will be a totally different spec that you need to pick before you leave town.
So, with hopefully newfound understanding of what the matchup chart tells us, here's the key:
In a siege, the goal of 250 people is to win the siege against the other 250 people. That's the objective.
This often gets broken down into sub-tasks, in the same way that in order to win a match of smash melee, you have to win sub-tasks (take 4 stocks). Then to actually take a stock you have to accomplish sub tasks like "control the stage" and "outspace them" and "defend properly" and all of that.
In Ashes, in order to actually "win the siege", that gets broken down into sub-tasks like "defend the gates" and "destroy the ballista", and "kill the dragon for the buff", etc.
Whether or not a particular match at all follows the optimal matchup theory is entirely a different story.