Some bitching about ELO

[pep]

burn showoff gg, I remember capping at 1400 few version ago.

[Stealer]

So having 200 more elo than the other team is basically 75% win chance so seems logical.


http://www.bobnewell.net/nucleus/bnewell.php?itemid=279
( plug in whatever elo you get curious about )

[Burn]

Do we use the same elo system though?

-Burn

[Stealer]

Uh yes?

Do you even elo?


Elo is a specific rating system.

[supersexyy]

Uh yes?

Do you even elo?


Elo is a specific rating system.

That doesn't mean that the numbers used in ID are the same. For example, I took the last game of ID: https://entgaming.net/customstats/islan ... e/8105896/

Loser lost -13.9 and winners gained +2.78. Obviously not reflective of the chess elo rating system.

[supersexyy]

delete

[Burn]

Exactly what I'm trying to say. Our elo equation is modified. Therefore, Stealer's link isn't accurate.

Do you even elo?

I also play chess, so I Elo very much.

-Burn

[Stealer]

That doesn't mean that the numbers used in ID are the same. For example, I took the last game of ID: https://entgaming.net/customstats/islan ... e/8105896/

Loser lost -13.9 and winners gained +2.78. Obviously not reflective of the chess elo rating system.

And? Different K values will result in different gains.

I also play chess, so I Elo very much.

uakf stole the formula from chess hence his chess relate comments...

/* FIXME: this is the chess distribution; games should be
able to set their own. */
if (player_ratings[i] < 2000)
K = 30.0;
else if (player_ratings[i] > 2400)
K = 10.0;
else
K = 130.0 - player_ratings[i] / 20.0;

tbh arguing with you people is useless since you probably can't even do "complex" math anymore nor do you try to actually look up any of these formulas to show them as different.

[supersexyy]

Taking the example of the game I linked:

Your rating: 1023 (average of builders) Opponent's rating: 981

Probability of winning: 0.560150145231903

So there should be a 44:56 elo gain/elo loss ratio

The ratio was actually 13.9 : 2.78 which is an expected 83% win ratio.

To get that expected win ratio you'd be looking at approximately 1250 elo builders vs 981 Titan elo.

@Stealer Happy for you to show me exactly what formula is used in ID and how it is reflective of the win% chances of the elo formula.

tbh arguing with you people is useless since you probably can't even do "complex" math anymore nor do you try to actually look up any of these formulas to show them as different.

An ad hominem response doesn't make your argument any more valid.

[supersexyy]

delete

[Stealer]

So there should be a 44:56 elo gain/elo loss ratio

no.

An ad hominem response doesn't make your argument any more valid.

Sure but I have no interest in argueing with people who can't keep up with the math.

[supersexyy]

Correct me then, instead of being a dick about it.

[Burn]

Chess elo calculations and ID calculations are CLEARLY different. You don't even need to check the elo equation if you pay attention to both chess and ID elo gains and losses.

At this point your ad hominem arguments are making you come off as a troll, so you're right, there's no point arguing.

-Burn