| |
BGonline.org Forums
Coming back to the OP. Luck is a major factor
Posted By: Michael In Response To: Answers @Bob, Dmitriy, Michael, Sebalotek: PR and luck in live matches (Florin Popa)
Date: Thursday, 21 September 2017, at 8:13 p.m.
Thanks Florin for the info. Leaving the weaknesses of the PR system aside,
I think Jeom's idea is actually implementable, but it can't be done by removing luck from the results. Because Luck itself is usually much more than the Total Error of the Players.
Let's take an example from a recent match where player A PR was 6 and B's 4 in a 19pt match: Player A Total Error EMG -3.8 Player B Total Error EMG -3.1 Player A Luck EMG +3.4 Player B Luck EMG -3.25 Summing up Total error and Luck we get Player A -0.4 Player B:-6.25 No wonder Player A won this match. It's not unusual both player to have +ve luck or both -ve. I haven't found a case that a player had won while having less luck than the opponent and also made more errors. There could be some, I am not sure, but I haven't found any.
Notice the effect of luck. In this particular case it was 9.5 times more than the Play error difference between the players. Such Luck EMGs are very common so Luck is more often than not the deciding factor of who the winner would be. So the first things in starting Joem's project imo, should be
a)to collect a reasonable number of matches (>1000) of a specific Length and Tabulate Total Error and Total Luck of the Winner and Loser. b)check if it's a rule that Sum of Error+luck EMG Winner >Loser c)Tabulate PRs d)scrap some matches so that Grand Total Luck of both players becomes near zero
After getting this data the rest should be easy.
NB. I think both luck and EMG don't depend on the match length, but it's better to start the project with a specific match length just in case. EMG is reported by bots as the Total error in a match, whereas mEMG as the rate of Error.
| |
BGonline.org Forums is maintained by Stick with WebBBS 5.12.