Replies: 7 comments
|
Hi, thank you for your comments on the project. About your requests: Technically, this could of course been done easily. Initially, I left it out intentionally, as I didn't want to put focus on a ranking with its competition for the first places, but rather on each engines personal progress. I felt (and feel) that there are already some very good lists for the top engines, so I thought that adding to that would not be necessary. So my list is rather for smaller engines that want to keep track of their progress. I also felt that giving ranks and places on the list would require an explicit policy how to deal with sf clones, derivates, AI created or assisted engines, and so on. On these topics, I lack the expertise and knowledge to judge over others work, so I stayed away from it completely, and leave these judgements to others with more experience. That is also part of my decision not to give ranks. I still did order my list after VLTC ratings, which implies such ranks in a way. Maybe I should switch to alphabetical order, or otherwise be honest to myself and add the ranking numbers. I'll think about it. |
|
-Ability to sort by various columns, name, release date, different time control Elos. I have added this sort capability to the list. You should also be able to filter the list when looking for a certain engine. |
|
-Error margins for ratings The error margins were already there, in the engine detail pages in markdown format. I linked these pages now in the version column, in order to make them accessible directly. |
|
-Some more information as well would be nice, I couldn't see what time controls or settings (like hash size) are actually being used I have added the testing conditions to the bottom of the list. |
|
-Specific matchup results like on CCRL, transparency is always a good thing I'll need some thinking on that, because this data is currently not in the database, and I have no idea how to excerpt it directly from the pgn-file. Happy for any hints, if there is a quick way to do it. |
|
@computer-chess-index |
|
I now have added a second complete list: https://computer-chess-index.github.io/cci/CCIRatings2.html The change towards one line per TimeControl now allows adding the additional statistics, so they are directly visible now and not only hidden in the detail pages. With the previously added search/filter options there are many new possibilities, like including previous versions also, so there will be even more soon. Any feedback / ideas for further improvement is welcome. |
Uh oh!
There was an error while loading. Please reload this page.
Hi,
I just came across this rating list today and I think it looks like a nice initiative. I like how thorough this list is with version names and dates. I also really like the novel ideas of basing off a random mover = 0 (which I am presuming based on the rating charts), and all time controls being in the mix together is something I've never seen before but makes sense and I like it, just make sure ponder is off.
I have a few requests for UI:
-A ranking number at the left of the table
-Ability to sort by various columns, name, release date, different time control Elos.
-Error margins for ratings
-Specific matchup results like on CCRL, transparency is always a good thing
-Some more information as well would be nice, I couldn't see what time controls or settings (like hash size) are actually being used
I realise this is github pages but these things would go a long way.
Validity of rating lists:
I saw someone saying they thought this rating list was faked. That seems far fetched to me but of course I don't know, releasing pgns of games would provide some validity to your ratings.
Good luck with your list and I hope to see it continue to improve.
All reactions