about team balance for tournaments

[vinme]

i'm thinking lets make it mandatory for every participant to get a Vrating and then just set up the balance based on that. 

if we subtract 1000 from the players rating (example 1600-1000=600) and divide that by 100 (600/100=6) we can get the points worth for the player.

cant get negative scores so anyone below 1000 will be a 0 point player

the issues though are the following: 

1:what would you want to be the average rating of all players per team?

2:the ratings aren't necessarily linearly balanced.

do for example

2 1500's = 1 2000 ?

2 1100= 1 1200 ?

3 1300= 1 1900 ?

 

tbh it often works good enough now that i look at it.

main factors to consider are:

1:what to subtract from the base Vrating to make ratings linearly balanced and if its not possible to do this although i'm guessing it is then we should use some other formula. like for example 1000-1300 each 100 ratings = 2 points  and 1600-2500 each 100 rating= 2 points .something like this might work.

2: what do you want an average rating of a tournament team to be? 1500? 1400? 1300? i'm guessing its best to have something average even low enough so many "good" players cant get together so that everyone can more easily get included so that teams will have to have some weaker players as well. if you think this is a horrible idea as then tournament will be filled with unskilled players and good players will be then id highly recommend setting the average to 1500-1700.

as you can see math can be done.it'll be simple math.everything seems to fit together decently after we decide.for example if we want a team average of a 1500 and we have decided that the ratings are linearly balanced then we can multiply 4 by the number of points that 1500 converts to(lets say -1000 was accurate as well) so (1500-1000)/100=5  and 5*4=20 this gives us the total points per team. for 3v3s and 2v2s we will have to divide total points in half or multiply by 3/4 to get the 2v2 and 3v3 TPA 15 and 10 respectively.

[vinme]

i was thinking 1000 is a good level to cut of from the rating as anyone below 1000 doesn't even know how the game works.but maybe we should go even lower

i think there isn't enough difference in points tho maybe for 1000-1200 compared to each other so 2 1100 would easily beat a 1200 i'm assuming so maybe we should cut off a bit less so like 800 so that way does this work?

so lets compare same(in this case an 1100=3 points 1200=4)

i'm very unfamiliar with lower rated player levels in relation to each other especially

does 4 1100 = 3 1200 ?

i think it does i guess

i'm guessing most optimal level is 800-1000 cutoff best.

if we don't cut off anything at all then there's an obvious problem like stockfish doesn't=2 1000 s 

def later on points represent the actual relevant ratings

i was thinking maybe we add a flat point so every player starts with 1 point and then they get added based on original 1000type points system.for example borg(i'm guessing Vrating will be 2300) would get 14 points(out of 24 as TPA will rise by 4 if this is added) as 1+ (2300-1000)=14

this would fix the issue with ratings just being nonlinear at the very least at the beggining and fix the graph.

def questionable if 1900 = 1300 for example id bet on the 1900 and with this system 1300s are worth more. so maybe you get 1 point (or more) flat at start and every 100 ratings after a certain high level point like 1800= 2 or something

[vinme]

7 minutes ago, go2die said:

@vinme full concept you wrote is wrong...  ELO rating works in 1vs1 somehow, and you want to reuse it for team games. 

You need introduce different approach based on different variables - economy, army power, kill ratio and others..  otherwise you can keep current rating it wont change anything much.

1v1 can properly show a players actual SKILL in 0ad as it shows how good a player is at 0ad period .eco military endurance and every other vital but irrelevant in in of itself stat perfectly balanced out by this magical formula that we call actually playing the fucing game and seeing results.

it does not show the players efficiency on factors that only come in team games such as but not limited to:

1.being a team player(giving res for various reasons,communicating,coordinating attacks ect)

2.the map being different in team games the edge players are much closer for example in tgs(played mainland normal) compared to standard mainland played small(or other maps that r becoming more mainstream for 1v1 like wrench,cross gr8 maps btw)that have similar distances to small mainland and the inside players being i guess even further from closest enemies than a med size mainland 1v1)

3.the game mechanic completely changes in a TG as you cannot as successfully counter a rush by making towers,making more men as enemy can just leave and attack your adjacent ally leaving you wasting a lot of res your and your teams only option is to counterattack that enemy roughly speaking.since edge players are close lower distance= higher incentive for early on aggression.

4.cav have a lot more value in tgs as mobility= you can move to target and change targets much quicker).cavrushes have more power in tgs as you can change targets if one spends enough on prevention.

making a "tg elo" is very difficult because stats don't necessarily correlate with anything.someone might have 1/5KD but it only happened because ally was taking all the beating.this obvious issues are very well understood by all of the players.i think i read some thread where someone  made a tg elo it took bunch of tgs and used some formula i cant recall. i suppose averaging out tg wins/losses and also taking into account total team power based on something like my formula can somehow show someones tg skill as you need another factor(actual rating being a great one) to in relation to it show tg skill.

either way we shouldn't wanna balance teams based on a tg skill taken into account as that's what we want the competition to be in i'm hoping.the whole point of this is to just make sure no op teams are formed by the factor of just brute skill and to encourage anything that affects team skill to lead to success.if we take into account tg skill when balancing then we are punishing the tg skilled players.yes we want even(er) matches but we also want a goal for all the teams to strive for and for a winner to have VALUE as they have accomplished something that isn't arbitrary this something being their superior skill in team playing(not in pure 0ad skill which i'm trying to eliminate with these methods i'm proposing)

 

you are not seeing the full picture or even the point of what i'm trying to do.

btw tell me why my concept is wrong.

id wanna take out actual 0ad skill out of the equation to make this a "team tournament" that rewards best "teams" in team skill.