Hi Mislav!
I really don't think cherry picking is as big a problem as I imagine you do. To me the biggest problem is collecting matches from live tournaments and worrying about cherry picking comes a long way down on my list.
If you don't trust the PR values you have an option to increase the threshold of the number of matches that a player needs to have played in order to enter the list. Say you don't trust an average unless it is at least 50 matches:
This one is perhaps more reliable than a list where 10+ is needed to be included in the list.
When it comes to collecting this information it is a bit of a problem since often the data is very lacking in the XG files that I receive:
Who transcribed it? The one who sent me the match? Sometimes yes sometimes maybe not. I also pick up matches from other locations: dropbox, othellos database ++ so that information may not be available either. For this information to be worth collecting more players need to include this when transcribing the match:
But sharing the match may not have been on the mind of this player when transcribing it and adding it to tens or even hundreds of matches when they decide to share the match may not be a top priority. This is also a reminder that it is nice to include transcriber information into the XG file so this information can be obtained later. XG remembers the 'Transcriber' field between transcriptions so you only have to enter your name once.
I don't think collecting more uncertain data will make the original data more certain. Then again it is easy to add support for this:
The Transcriber is automatically set if it is present in the XG data. As for the 7777 existing matches. Do you volunteer to go through all of them and update this information? If you do I can give you access.
If you do see someone who are recording a lot of matches but you only see a handful of matches in the database (all of them very well played) you might start to suspect foul play but should hire an army of spies who will take a note of everyone who records their own matches? I don't think so! :-)
I don't think I will do anything in particular with this information so it is just an added info entry on a match for now.
Backgammon Studio The Swiss Army knife of backgammon!
Best regards,
TP
PS. Don't hesitate to provide feedback on any issue you like! Some feedback is better than no feedback! :-)