I am very interested to see what impact the introduction of CSW has on gameplay, so I did a little statistical study a few days ago – posting the results today.
I did the study on 15 local players who play rather regularly here – and I excluded all non-members (because it is not easy to access their averages and tourney records – Sorry Ben and Tony), and the results can be found here.
This is how I got the statistics:
1) For OSWL averages, I calculated each player’s average from 2006 till the JGCC Mini in April 2007, which was the tournament before the introduction of CSW.
2) For CSW averages, I calculated each player’s average from the Singapore Open till the end of 2008.
3) Only locally rated Division A tournaments are counted (because the standard of the field is the most consistent there) – this means I excluded Causeway tourneys and all Division B averages.
Some factors have to be taken into account:
1) Some players have improved drastically from 2006 and hence their increased averages will not reflect the dictionary’s impact on scoring. Because of this, I have done an additional calculation – each player’s average rating difference from OSWL to CSW. The start points and endpoints for the rating calculation are the same as the average calculation. I calculated average rating by averaging each player’s ratings after all tournaments. The stats can be found here. Note that the calculations may not be as accurate as the former one because I was multitasking when I did this :). I have not come up with an algorithm to incorporate these figures into the earlier calculations, maybe someone can help me with that.
2) A corollary from 1) is that some players’ averages will invariably drop to counteract the relatively larger gain in averages by the improving players.
3) Although I have tried my best to calculate averages against similar fields, the “field” for each player might be different, especially in majors where the standards tend to vary more widely – so a player at the bottom might be getting higher averages against people lower down, whereas he/she’d not be able to do so against people higher up the table.
4) Since the field for each tournament varies (rather invariably heh heh), and not everyone plays in every tournament, people will again get different “fields”. This is probably not such a major problem though because I have calculated the averages over (I hope) a long enough period of time.
Based on the figures in the document I have estimated (rather arbitrarily) that the new dictionary has increased game averages of about 4-6 points, excluding player improvements. Perhaps a better gauge can be done by pitting Quackles against each other (a pair consisting of two Quackles with the old lexicon, another comprising two with the new lexicon) and seeing how the averages differ over say 100 games for each pair.