How exactly do we expect low rated players to play the game?
-
Even read in Discord recently from some ~2k player (was it zwaffel??) that being matched together with a 1k against 2 ~1600 is just unwinnable.
If it's impossible for a ~3000k team to win against a ~3200k then the score isn't a good predictor of team composite performance. Do you understand that?
-
@clyf that is not the fault of the rating system per se, but more of the different environments the players play in. because if they'd play more games together then the ratings would start to match their real skill for that gamemode again. It's the same how some players get to 3.6k rating.
Rating only really makes sense if there is sufficient exchange between the whole playerbase.
And of course the high volatility in the performance of a low game count 1k player. might work well might not and also 50% winrate is considered an autoloss for the average player since for a game to feel balance you need a 66% winrate, which is impossible for everyone.
But in this game the low rated player has a high impact on the game, if it's actually unwinnable is something that can and should be questioned. -
@nex said in How exactly do we expect low rated players to play the game?:
No idea what they currently use for team slayer or what trueskill 2 is
Trueskill 2 is a proposed system that includes more metrics (like kills) than just win/loss/draw in its statistics considered.
(unlike trueskill)
which paper you read might be very important here, since trueskill (1) has been released, and assigns skill based ONLY on whether a game was won, lost, or drawn (no KDA or ingame score or frags or anything).
Trueskill 2, I believe, still has its maths 'hidden' (and all the overhyping that such a move enables!) This might have changed since I last read about it a few years back, though.As I understand it, trueskill 2 is geared towards FPS games with scores, which is why I thought it might be what was being referenced here regarding 'teamslayer modes' and the like. At the very least, it's confirmed to take other factors besides a binary victory/defeat into account.
-
that is not the fault of the rating system per se, but more of the different environments the players play in
If there's one environment where the player is rated at 1000 and winning 50% of their games (steady state, 4v4 with 4*1k/4*1k rated players in game), and another where a player is rated at 1000 and getting crushed (2k,1k/2*1.5k), and no further distinction is made between those two environments, then their score is converging on a different point for each environment, and no amount of games played will result in their score converging on a single point.
(emphasis for emphasis, not to be a dick)
-
@clyf my point was not that the statement "2k/1k vs 1.5k/1.5k is unwinnable" is true, but that the opposite of your statement "low ranked players have no influence on the match" also exists and the truth lies somewhere in between.
The reason most 2k+ players don't play with <1.5k players together is that it immediately ruins the game, which would not be the case if they had no impact. Lower rated players are just a lot more inconsistent than high level players, which makes a game more of a coin flip. -
My original statement in regards to how TrueSkill evaluates expected team performance:
At higher level games, where players have the knowledge and wherewithal to identify and exploit weaknesses, it seems the weak link is the deciding factor.
At lower level games, where players are on average less able to identify and exploit weaknesses, it seems the strong link is.
To refine the entire point: TrueSkill assumes that the relationship between individual performance and composite term performance is mathematically linear, while empirical evidence in FAF suggests that it is not.
-
Okay, here's our problem--
I said:
your kills and deaths being reflected directly in the game deciding score
You said:
I can tell you that is wrong. The only relevant metric is you won
What you thought I was saying was:
your kills and deaths are used to calculate your TrueSkill rating
... whereas what I actually meant is:
Your kills and deaths [are summed with the kills and deaths of your teammates] to determine the score [which determines who wins the game, which is then used to calculate TrueSkill]
At no point did I say that any metric other than win/lose was used in the TrueSkill calculation.
-
@clyf said in How exactly do we expect low rated players to play the game?:
At no point did I say that any metric other than win/lose was used in the TrueSkill calculation.
Yes, but you suggest that this is a metric that shows how good you are / how much you contributed to the win, which is false as there are parts to the game that are not reflected in that score. (like killing an enemy about to kill your teammate, which in turn generates more kills)
@clyf said in How exactly do we expect low rated players to play the game?:
To refine the entire point: TrueSkill assumes that the relationship between individual performance and composite term performance is mathematically linear, while empirical evidence in FAF suggests that it is not.
As soon as teamplay is involved the performance of the team is no longer simply the sum of the individuals, otherwise it wouldn't ever make a difference if you play with a friend vs. you playing with random people on your team.
But assuming a linear relationship is fine as it evens out with enough games, unless someone intentionally plays with higher rated players and takes the spot with the least influence on the game, but that's a problem of custom games in general. Sometimes your spot makes the difference sometimes it doesn't. Sometimes you are even the high rated player in a lobby.The rating system makes a statement about the statistical distribution of wins/draws/losses over time. So a game can feel absolutely unwinabble if you play 1k/2k vs 1.5/1.5, but that is just part of the randomness involved and you get that feeling because suddenly your own contribution is lower than you expect since a lot hinges on the fact if the worse player has a good day or not.
-
metric that shows how good you are / how much you contributed to the win, which is false as there are parts to the game that are not reflected in that score
You don't understand what a metric is.
Are you Armistice840 on the discord?
-
How the hell did this thread veer so far off course? Shoo.
-
I wanted to downvote you but I'm going to make the right choice and bail on this whole conversation instead.
-
@clyf I get what you're saying.
Many teamgame systems, on top of just adding up the total skill of each team, also add modifiers for various factors (I'm sure you know this, but I'm just detailing what I'm talking about before exploring further)... Eg, modifying the calculated 'rating' for a team, both for matchmaking and post-game skill adjustments, for a large rating spread (1000+3000 vs 2000+2000).
Another common adjustment is for pre-made teams (huge* in MOBA games).I understand your (Clyf) point about how such modifiers might be harder given the nature of high-skill vs low-skill games, and how a large rating spread can be an advantage in one place, and a disadvantage in another; but I don't think it's impossible to overcome this, by plugging in more variables... Eg. make the adjustment based on both the spread, and the average rating, or even factor in maximum or minimum rating.
I don't envy whoever tries to do it though! I'd expect people to have vastly different opinions on where to start, and I'm not sure whether there's enough data to find a perfect 'maths only' approach!I assumed FaF already used these kind of modifiers, but I admit that was perhaps a bad assumption to make! It might be relevant to ask a developer about it?
-
@clyf said in How exactly do we expect low rated players to play the game?:
Are you Armistice840 on the discord?
I take that as an insult.
But anyway
@clyf said in How exactly do we expect low rated players to play the game?:
I wanted to downvote you but I'm going to make the right choice and bail on this whole conversation instead.
Guess we can just agree to disagree on the performance of the rating system.
@redx said in How exactly do we expect low rated players to play the game?:
How the hell did this thread veer so far off course? Shoo.
Not sure if we actually went that far off
The whole discussion about the rating system was, because the reason for higher rated player to have rating requirements on their game is because (they think) mixing in with lower rated players destroys their experience. So we were discussing if this was a problem with the rating system (so a 2v2 of 1k/2k vs 1.5k/1.5 is truly just unbalanced) or if this is just because lower rated players inherently play more unreliable.
If this was simply a problem in the rating system it could be changed. Then games with higher rating discrepancies would be more balanced and thus more fun.@sylph_ said in How exactly do we expect low rated players to play the game?:
make the adjustment based on both the spread, and the average rating, or even factor in maximum or minimum rating.
I don't envy whoever tries to do it though! I'd expect people to have vastly different opinions on where to start, and I'm not sure whether there's enough data to find a perfect 'maths only' approach!Yeah while this sounds like a nice idea, I don't think there will be a mathematical solution that is consistent in it's predictions (even less so, when you go the step further and try to incorporate map slot influence).
@sylph_ said in How exactly do we expect low rated players to play the game?:
I assumed FaF already used these kind of modifiers, but I admit that was perhaps a bad assumption to make! It might be relevant to ask a developer about it?
I think there is a premade bonus for TMM, but nothing aside from that I think.
-
I apologize, it was a low question.
I agree I don't think we strayed too far from the main topic.
A metric is a quantifiable measure of the system, but in practice does not (and cannot) account for every element of a system. Win/lose, kills:deaths, the number given by the TrueSkill system--none of these account for everything that can happen, yet all are metrics.Instead of belaboring this point further, I think the way forward is to experiment with either A. introducing additional metrics into the score calculation a la TrueSkill 2 or B. modifying the score calculation function to something other than a linear sum. Let's conclude here until we have something to discuss in that regard.
-
@clyf said in How exactly do we expect low rated players to play the game?:
A. introducing additional metrics into the score calculation a la TrueSkill 2
I think that's off the table (ftx also often rants about how this would be a bad idea), because players play the rating system and whichever metric you use to calculate their rating, the players will try to maximize that and any metric aside from win/loss will inevitably warp the game and might even cause toxicity within the team. (kills/eco/quick win/score, all of these will lead to certain kinds of abuse)
@clyf said in How exactly do we expect low rated players to play the game?:
B. modifying the score calculation function to something other than a linear sum
While good in theory (as I already mentioned in my response to sylph_), I don't believe there is a mathematically sound solution to this, as it is very opinion based.
I think the problem is also that in custom games the your contribution-rating ratio is very "random" since there is no control how/where you got that rating and how it compares to what you are playing now.
And in ladder the sample sizes are quite low, since there aren't that many games played and almost no high level players queue ladder/tmm.Sidenote:
@clyf said in How exactly do we expect low rated players to play the game?:A metric is a quantifiable measure of the system, but in practice does not (and cannot) account for every element of a system. Win/lose, kills:deaths, the number given by the TrueSkill system--none of these account for everything that can happen, yet all are metrics.
I guess me calling these metrics is wrong
Mathematically the metric we want is the players ability to play the game and cooperate with their team well, because that directly influences the games outcome.
But we can't measure that directly so we approximate this by using other "metrics" and any "metric" where A > B (in the skill metric) and A <= B (in the the approximation metric) is just a wrong metric to me.
For win/loss you could probably argue that this is also the case for certain games, but it will average out given enough games. So a player that consistently wins more games than he looses is better than his opponents were, while a player that consistenly has a high kda is not necessarily better than his opponents were.
(Just so you understand where I'm coming from. We should probably cut this discussion here and accept we have different definitions/assumptions) -
Systems that revolve around in game metrics are used, but the firms that use them specifically blackbox the exact calculations and people just guess what the potential metrics could be/are weighted as. This isn’t exactly an option in an open source project and so you’re immediately going to have people making maps that give them a bias towards gameplay that boosts their rating or playing in a way that does it.
Too much player freedom on FAF to ever allow metric based ratings.
-
I don't know, kind of feels like we're exiting the vampire castle.
A rigorous analytical solution for a better team rating is beyond my mathematical ability without some heavy research, but that doesn't mean there isn't one floating around out there.
Also, given the setup we have, it's also possible to silently test alternative systems (TrueSkill was tested in the same way against Elo, graded based on their percent of wins predicted correctly).
Mathematically the metric we want is the players ability to play the game and cooperate with their team well
I would refer to that as a goal/outcome--but now I know where you're coming from. And I agree, that is the outcome I'm looking for.
I can spitball about what metrics would be good for FAF (in game score? mass efficiency? kills?) but can't confidently say that using any of them wouldn't be abused/go sideways.
A meta-level correction would be to incorporate map statistics. Map gen maps have a larger effect, maps that other people play a lot have a larger effect (and vice versa), playing the same map over and over has a smaller effect*.
*Could grade this based on size, terrain, and distribution of mass spots to catch all the gap clones.
EDIT: After asking my friend Mr. GPT-4, looks like TrueSkill2 is a good starting point for all the above.
