| |
BGonline.org Forums
title hierarchy
Posted By: Henrik Bukkjaer In Response To: title hierarchy (Stein Kulseth)
Date: Sunday, 3 April 2011, at 8:26 p.m.
I've learned my statistics in Danish so excuse me if the following is incorrect translated.
Correlation does not imply causality - at least I didn't mean to. There could very well be a spurious causality here.
Our observations was simply, that the lower number of players scoring TMP over 1200 rating points when you go back, pretty much matches the number of players and games played at what we today would call the highest level. The strongest players have indeed improved their game mainly due to the bots. And for the players just under world-class it's the same.
I know that one thing is absolute (though slightly objective) and the other is relative (ratings), and as such are like comparing apples to oranges.
But consider a population of players in a rating system at a stable state (if that exists in backgammon - should be read as enough TMP turn over to get the proper distributions). Now you introduce a strong tool to improve your game (bots, at a certain level, or improved bots). Some players will use them study harder, etc. Others will continue to play their same game. The improved playing strength of the best players would indeed give impact on these players ratings compared to the other players - it would stretch the population. Some of it might also be simple coincidence. I have some charts I did when analyzing some of the trends for different thresholds I should show you there was some fun patterns: First a big raise in the percentage of matches over 1200 and 1250, then followed by a continued raise in the 1200 "revenue" but a drop in the very top ratings "showing us" that the relatively difference between the top-players has evened out though the top-players compared to the "normal" players have bigger difference than before. This could very well be driven (or amplified) by strong bots.
Well it's getting slightly OT now :-) But just to tell you, that we did spend time to analyze the evolution in ratings/TMP over time, to analyze if we should adjust (lower) the requirements for the norm, going back in the historic data (to hand out norms at a lower threshold to players that was strong and active 10 years back or more). We decided not to. Simply because we would then honer the very strongest and most dominant ones from back then, yet keep the level at something closer to today's "1200+" standard. A good trade-off between old school titles and not flooding the system with titles for players no longer at the top of the game.
:-)
| |
BGonline.org Forums is maintained by Stick with WebBBS 5.12.