Code center > Darwinbots3
DB3 Leagues
Panda:
We don't know the elo of each of the 10000 players when we're trying to simulate that, unless we just assume they're all new?
Peter:
There's need of some boilerplate code to keep track of all players.
Btw, I'm curious how well Elo/TrueSkill/Glicko/win% compare when matches are picked at random.
Panda:
Yeah, there is boilerplate code for it but I'm just trying to work out how to do it.
Each of these systems are designed so that matches are picked at random, aren't they? You basically want to take a small sample that represents the global population.
Peter:
--- Quote from: Panda on April 12, 2015, 11:51:14 AM ---Each of these systems are designed so that matches are picked at random, aren't they? You basically want to take a small sample that represents the global population.
--- End quote ---
They're not. Otherwise a chess grandmaster has to play a low ranked player often. They're designed to calculate the right strength of players. Including on how to deal with matches not being a random sample.
Edit: I was playing around with TrueSkill. And the library I took even has issues with high vs. low ranked players. Calculating skill as NaN in some cases...
edit2:
--- Code: ---using System;
using System.Collections.Generic;
using Moserware.Skills;
using System.Diagnostics;
namespace TrueSkillTest
{
class Program
{
static void Main(string[] args)
{
GameInfo defaultInfo = GameInfo.DefaultGameInfo;
List<BotPlayer> botPlayers = new List<BotPlayer>();
for (int i = 0; i < 10; i++)
{
var newb = new BotPlayer(i);
newb.rating = defaultInfo.DefaultRating;
botPlayers.Add(newb);
}
var random = new Random(1234);
for (int i = 0; i < 250; i++)
{
int id1 = random.Next(10);
int id2 = random.Next(10);
if (id1 == id2)
continue;
if (id1>id2)
{
int switchId = id1;
id1=id2;
id2=switchId;
}
var team1 = new Team(botPlayers[id1], botPlayers[id1].rating);
var team2 = new Team(botPlayers[id2], botPlayers[id2].rating);
var teams = Teams.Concat(team1, team2);
var results = TrueSkillCalculator.CalculateNewRatings(defaultInfo, teams,1,2);
if (results[botPlayers[id2]].ConservativeRating.Equals(double.NaN) || results[botPlayers[id1]].ConservativeRating.Equals(double.NaN))
{
Debug.WriteLine("NaN happened "+id1 +" "+id2);
continue;
}
botPlayers[id2].rating = results[botPlayers[id2]];
botPlayers[id1].rating = results[botPlayers[id1]];
botPlayers[id1].wins++;
botPlayers[id2].losses++;
botPlayers[id1].games++;
botPlayers[id2].games++;
}
Debug.WriteLine( "id \t ConservativeRating \t Mean \t\t StandardDeviation \t win% ");
for (int i = 0; i < 10; i++)
{
var rat = botPlayers[i].rating;
Debug.WriteLine(i + "\t" + rat.ConservativeRating + "\t" + rat.Mean + "\t" + rat.StandardDeviation + "\t" + ((float)botPlayers[i].wins / (float) botPlayers[i].games)*100 +"% ");
}
}
}
public class BotPlayer : Player
{
public Rating rating;
public int games, wins, losses;
public BotPlayer(int i)
:base(i)
{
}
}
}
--- End code ---
TrueSkill ratings and win% after 250 random matches.
Player with id 0>1>2>3 etc.
id ConservativeRating Mean StandardDeviation win%
0 36.6018093613847 44.3298552173074 2.57601528530759 100%
1 32.3380042534196 39.2166471449276 2.29288096383602 91.42857%
2 32.2120554191209 38.422586286346 2.07017695574173 89.58334%
3 24.4339328188783 29.6447006659815 1.73692261570105 50.9434%
4 23.2533477507645 29.1590048994883 1.96855238290791 68.18182%
5 18.3016520660433 23.4776208283092 1.72532292075527 44.64286%
6 13.2212442413914 19.1699259868905 1.98289391516637 26.82927%
7 10.0415863402407 15.7679791596343 1.90879760646453 25%
8 4.18826635383718 10.6105575553436 2.14076373383547 10.41667%
9 -2.27059972504378 4.86224661720219 2.37761544741532 0%
edit: RPC, TrueSkill same amount of matches
Major strategy is Scissor. Paper and Rock as minor strategies. As you can see the minor strategy beating the major tactic does take place.
Rock Paper Scissor
id ConservativeRating Mean StandardDeviation win% games
0 22.2911857672074 25.635469626594 1.11476128646221 29.54545% 44 Scissor
1 21.7841748039772 25.0038555837073 1.07322692657669 17.77778% 45 Scissor
2 19.0423325347074 22.3747330556796 1.11080017365737 26.92308% 52 Paper
3 24.8168168325261 28.5840867091271 1.25575662553368 68.88889% 45 Rock
4 22.2261144113132 25.2695586097977 1.01448139949485 23.07692% 52 Scissor
5 21.3862248458399 24.6455366523104 1.08643726882351 15.55556% 45 Scissor
6 20.652998113402 24.0505831870944 1.13252835789746 11.62791% 43 Scissor
7 17.4614961926684 21.3511453077313 1.29654970502095 20% 40 Paper
8 22.4296910660544 25.6949142849584 1.08840773963469 34% 50 Scissor
9 24.2526401114872 28.2686183545096 1.3386594143408 71.42857% 42 Rock
Numsgil:
--- Quote from: Peter on April 12, 2015, 01:10:18 PM ---edit: RPC, TrueSkill same amount of matches
Major strategy is Scissor. Paper and Rock as minor strategies. As you can see the minor strategy beating the major tactic does take place.
Rock Paper Scissor
id ConservativeRating Mean StandardDeviation win% games
0 22.2911857672074 25.635469626594 1.11476128646221 29.54545% 44 Scissor
1 21.7841748039772 25.0038555837073 1.07322692657669 17.77778% 45 Scissor
2 19.0423325347074 22.3747330556796 1.11080017365737 26.92308% 52 Paper
3 24.8168168325261 28.5840867091271 1.25575662553368 68.88889% 45 Rock
4 22.2261144113132 25.2695586097977 1.01448139949485 23.07692% 52 Scissor
5 21.3862248458399 24.6455366523104 1.08643726882351 15.55556% 45 Scissor
6 20.652998113402 24.0505831870944 1.13252835789746 11.62791% 43 Scissor
7 17.4614961926684 21.3511453077313 1.29654970502095 20% 40 Paper
8 22.4296910660544 25.6949142849584 1.08840773963469 34% 50 Scissor
9 24.2526401114872 28.2686183545096 1.3386594143408 71.42857% 42 Rock
--- End quote ---
Looks like it places them all around the same skill? I think that's what we'd want if so. It's hard to get a sense of the relative scales. Can you add in a strategy that has a 80% chance to win against any of RPS? I'd like to see if it's noticeably higher than all the others.
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version