4NCL Online

[MartinCarpenter]

I suspect that with juniors improving fast the best answer might be to test them at something like current grade + 200? If they're still showing as way out of line then you've likely got a problem.

Actually I'd test everyone at grade +100 as a matter of routine. Trying to compensate a bit for things like great tournaments.

Thomas' suggestions are fair enough, but you really should be testing on a big enough sample of games to take that sort of effect out of the equation. Which I realise might not be comfortable in terms of a single season of the 4NCL.

You definitely don't want successful appeals to be a routine thing.

[Pete Heaven]

In response to Roger's point about comparing 4NCL results with OTB play, it can be quite illuminating, or at least interesting. Taking the combined scores for the 3 Correspondence players that I thought might do well before the event began (all about the same rating), they have a combined score of +20 =6 -0 with an overall rating performance of 214 against an average opponent rated 171. I then estimated their rating performance on the equivalent number of their most recent OTB games and the average opponent was rated almost exactly the same but their performance in these 26 games (they scored 50%) was, perhaps unsurprisingly, 171. It proves nothing and, having analysed all 26 4NCL games I suspect any z figure would be well shy of 4; however I found myself, with 2 engines running, remarkably adept at predicting their next move.

[Roger de Coverly]

Some teams in that tournament video and audio record themselves playing so that if they were banned they could submit that as proof of their innocence.

That's an interesting idea. Is it because of a fear of false accusations?

Logically then, if a black box detects that someone appears to be playing too well for their rating, video and audio surveillance is then imposed.

[JustinHorton]

I'm genuinely intrigued how a good appeals process would work. In particular on what grounds could you hope to succeed?

Ok, I'll try to give some ideas. Let's say you've played only 5 games and 4 of those show it's very likely you were using a computer, on their own not enough but put together it doesn't look good for you.

For these games you can show:

a) You were aware of all/almost all the moves before the game. Perhaps from an article, online video, Chessable course or correspondance game (that might not have shown up as 'known theory'.

b) You had specific preparation for a certain game. Look, my opponent played these 17 moves already before and so I did some work with an engine. Perhaps you can even send over the file that you created before the game.

c) It's a position you've played a lot (either over the board or online) and all the ideas are well known to you.

d) Your opponent blundered early on and the moves were rather easy to find after that.

There may be other examples/reasons/excuses. Of course it's rather lucky (or unlucky given you've been flagged up for cheating) that these should occur in only 5 games, but it would cast reasonable doubt.

Below I give an example of c) as I've had this type of position after 8 moves many, many times. The game was played online at a 10+2 time limit. I'm sure most players could find individual examples of these things happening in their games.

1. e4 c5 2. Nc3 Nc6 3. Bb5 g6 4. Bxc6 bxc6 5. f4 Bg7 6. Nf3 Nf6 7. d3 O-O 8. O-O d6 9. Bd2 Qb6 10. Na4 Qb5 11. b3 Rb8 12. Qe1 Nd7 13. c4 Qb7 14. Bc3 e6 15. Bxg7 Kxg7 16. e5 d5 17. Qf2 d4 18. Ng5 Qc7 19. Qh4 h6 20. Ne4 Qd8 21. Qxd8 Rxd8 22. Naxc5 Nxc5 23. Nxc5 Bb7 24. Rae1 Bc8 25. g4 f6 26. exf6+ Kxf6 27. Re5 g5 28. Ne4+ Kg6 29. fxg5 hxg5 30. Rxg5+ Kh6 31. Rf6+ Kh7 32. Rf3 Kh6 33. Rh3# 1-0

While everything Thomas says here is reasonable (and apologies for quoting in full, I know it can make the page hard to read) it is very hard to prepare an appeal when you do not know the grounds on which you are accused.

[Thomas Rendle]

Thomas' suggestions are fair enough, but you really should be testing on a big enough sample of games to take that sort of effect out of the equation. Which I realise might not be comfortable in terms of a single season of the 4NCL.

You definitely don't want successful appeals to be a routine thing.

I think that's why it was generally waited until 8 or 9 games had been played. I believe at least one player is intending to appeal? Elsewhere I've seen various people claiming innocence on behalf of others, but generally with 'I know him, he wouldn't cheat - he's a strong players, etc.' rather than providing any evidence that I outlined above in A-D. Perhaps that will come in any appeal. Personally if I were accused it'd be the first thing I'd be looking to post, anywhere and everywhere, along with analysis of all the games.

[JustinHorton]

While everything Thomas says here is reasonable (and apologies for quoting in full, I know it can make the page hard to read) it is very hard to prepare an appeal when you do not know the grounds on which you are accused.

Also, you know, it's hard. It's hard because ir's a lot of work, and it's emotionally hard. When I put together this stuff (it's not an appeal as such, but it's the sort of thing you'd need to do to be thorough, and it is an appeal of sorts, to public opinion) it took me about six weeks.

[Peter Hornsby]

Some teams in that tournament video and audio record themselves playing so that if they were banned they could submit that as proof of their innocence.

That's an interesting idea. Is it because of a fear of false accusations?

Logically then, if a black box detects that someone appears to be playing too well for their rating, video and audio surveillance is then imposed.

Yes. False accusations were rife when we started, then we set a limit of appeals to the panel like they do with Hawkeye in Tennis or DRS in cricket and now they are almost non-existent. However anyone could always email [email protected] and report the player to Chess.com at any time.

That said it's a process of education (myself included), a lot of people believe(d) that 99% score on CAPS = cheating until we shared a data analysis of why this is not necessarily the case.

[Thomas Rendle]

While everything Thomas says here is reasonable (and apologies for quoting in full, I know it can make the page hard to read) it is very hard to prepare an appeal when you do not know the grounds on which you are accused.

I was under the impression the players (from the 4NCL) did know the grounds on which they were accused. They have not (as yet) chosen to publicise those grounds, neither have the 4NCL for understandable reasons.

[Roger de Coverly]

I was under the impression the players (from the 4NCL) did know the grounds on which they were accused.

There was a withdrawal of a team after a handful of rounds. Wasn't this triggered by a lichess ban on one of their players, rather than the 4NCL? lichess wouldn't give reasons, or at least not under conditions the 4NCL found acceptable.

[JustinHorton]

While everything Thomas says here is reasonable (and apologies for quoting in full, I know it can make the page hard to read) it is very hard to prepare an appeal when you do not know the grounds on which you are accused.

I was under the impression the players (from the 4NCL) did know the grounds on which they were accused.

I'd be interested to know, though, to what extent, and at what stage.

[Thomas Rendle]

Yes. And if I were accused (and were innocent) I'd want to post those along with an explanation. Perhaps we'll get this in due course.

[Matthew Turner]

I suspect that with juniors improving fast the best answer might be to test them at something like current grade + 200? If they're still showing as way out of line then you've likely got a problem.

Actually I'd test everyone at grade +100 as a matter of routine. Trying to compensate a bit for things like great tournaments.

Thomas' suggestions are fair enough, but you really should be testing on a big enough sample of games to take that sort of effect out of the equation. Which I realise might not be comfortable in terms of a single season of the 4NCL.

You definitely don't want successful appeals to be a routine thing.

Martin,
The calculations that Ken Regan does initially sets the ratings at 25 points higher than the actual rating. A further adjustment is made for lower rated players, but I am unclear as to what that is.
I do not believe that successful appeal will ever become a routine thing, but I'm very comfortable with having some successful appeals (and I think Lichess and Chess.com are too). In a sense, successful appeal highlights flaws in the system and show how things can be improved.

[MartinCarpenter]

I do worry a tiny bit about aiming for multiple appeals given how people seem to react so far. Its obviously quite an emotive thing.

I'm not surprised that Ken is being sensible 25 really isn't much leeway though, not when you've shifted the playing conditions from >4 hours FTF to somewhat quicker online from homes. Most of the population will obviously play slightly worse due to that.

I would also expect a small but real %age of the playing population to notably improve their absolute underlying playing strength.
(Only +100 or something, not +5/6!)

If you're being cautious I think you'd want to take that effect away. It shouldn't weaken the power of the statistics too much.

[Adam Raoof]

I do worry a tiny bit about aiming for multiple appeals given how people seem to react so far. Its obviously quite an emotive thing.

I'm not surprised that Ken is being sensible 25 really isn't much leeway though, not when you've shifted the playing conditions from >4 hours FTF to somewhat quicker online from homes. Most of the population will obviously play slightly worse due to that.

I would also expect a small but real %age of the playing population to notably improve their absolute underlying playing strength.
(Only +100 or something, not +5/6!)

If you're being cautious I think you'd want to take that effect away. It shouldn't weaken the power of the statistics too much.

Ken does adjust for time controls, for theory, for forced moves, for age.

[Thomas Rendle]

The calculations that Ken Regan does initially sets the ratings at 25 points higher than the actual rating. A further adjustment is made for lower rated players, but I am unclear as to what that is.

Just to be clear is that 25 ECF or ELO? 25 ECF is equivalent to ~200 ELO currently?