General Discussion

The latest login invited me to applaud Russell Sherwood's IM/SIM norm gained at the recent Dobri Semov Memorial tourney 2400-2499 https://www.iccf.com/event?id=103622 .

It went on to quote Russell's one win, courtesy of an inexplicable blunder by his opponent in a totally equal position - 25...Qd7??

In terms of Russell's ongoing work for ICCF as an administrator, the norm was as the author mentioned 'well-deserved'.

In terms of quality of play, it was manifestly not. It took me about two minutes to run through the entire game on SF 17 and detect the one and only error.

This has become a clear trend over the past two or three years in ICCF play.

Is this what CC and the ICCf has come to? Awarding meaningful titles on the basis of outright blunders [probably from boredom or loss of interest] or worse - points gained from withdrawals or even the very unfortunate physical demise of a player.

It fortifies the need to ban or police engines to keep the game alive, when the emperor is so obviously wearing no garments at all!

As if on cue, the two latest norms for Helen Sherwood and Paul Keevil [both for SIM] confirm the same trend.

Both Helen's win over Tom Biedermann https://www.iccf.com/game?id=1437581, and Paul's over Vladimir Antonov https://www.iccf.com/game?id=1455966 contain huge, game-losing blunders by the opponent - 29. Rd1?? in the first, 27.f4?? in the second - which are highly untypical of the player's rating. Both occur in totally equal positions.

The machinery of ICCF is currently deciding titles and championships by default. By means of inexplicable blunders of the kind which would be hard to understand in OTB games, let alone advanced engine-assisted corr chess!

Ratings and titles awarded on this basis in no wise represent the actual strength of the players involved, which in turn brings us back to the viability of engine-assisted chess at this level.

They simply highlight the futility of the game as it has evolved over the last recent years... Is it too late to remove engines from 'live' Corr chess. Can analytical engines [and the databases which accompany them] be confined to the research aspect BEFORE a game is underway, without reference to either during a game?

It would ofc require a huge effort on ICCF's part to begin policing the use of engines when Pandora's Box was opened over 20+ years ago now. But is it not an effort worth making?

The Corr game as it is is dying on its feet. If you do not blunder or withdraw from an event, chances are all your games will end in draws. Where is the joy in that?

Amici Sumus [but let us at least be 'wide-awake friends'].

Blunders in OTB games are (usually) punished by a win for the opponent, just as in CC. The engine is just a scapegoat in this discussion.... titles can be won or lost in either type of game. A more useful discussion might be how to avoid so many draws?

The situation is not comparable Paul.

OTB there are any amount of reasons for blunders - loss of concentration, physical or mental fatigue, poor tactical calculation, weakness under time pressure etc....

None of those apply in CC as it now exists as the engine is making 99% of the decisions and it is indefatigable.

So the real question in these games, and in several other recent Welsh-based games like them, is 'Why did the losing player involved suddenly ignore the engine's advice when they follow it for all the other moves in the game?' It's the kind of question which would attract the interest of fair play arbiters on a site like chess.com and rightly so. 

You yourself were probably asking it after your WCCC win over Fred Clough in only 16 moves!

Your question is strictly secondary to that. As long as you allow engines to be used the percentage of CC draws will continue to rise until they reach <100%. 

The only ways to avoid that without changing the character of the game radically are

1. Continue to encourage defaults/withdrawals/'blunders' as the only form of obtaining a positive result.
2. Restrict or ban the use of engines.

Chess.com's "engine detection" is nothing more than marketing snake oil... it can't be done reliably in correspondence chess.

If your paying customers believe you can do it, and you occasionally ban the odd sacrificial lamb, then the paying customers will keep coming back.

They're interested in your money, not in "fair play".

That's very presumptive on your part Austin. I'm quite sure there are ppl within chess.com who fully believe in the concept of 'fair play' and are happy to develop algorithms to improve it.

It's not perfect but at least they're trying and their model works.

Something like it is worth a try in ICCF if the game isn't to evaporate into an oblviion of predictability.

If you could show me a shred of evidence that "engine detection" can work reliably and consistently for high level correspondence chess, then I would agree with you... but there isn't any.

OK, suppose I win the first "No Engine ICCF World Championship" and someone says, "hey, that guy used an engine on move 43."... how do you handle that? - it's OK on chess.com where there is probably some wiggle room with sensitivity and specificity - but for an official world championship tournament?

Nope, it won't happen... and yes, it does seem quite likely that the game will evaporate into an oblivion of predictability - I think there is probably a general acceptance of that now, but it's only a hobby, there are plenty of others!

https://www.chess.com/article/view/chess-com-fair-play-and-cheat-detection .

"Having seen Chess.com's system in great detail, both the algorithms used and the 'team at work' (when I was on-site at the Chess.com Meetups), I can attest fully that Chess.com's approach is advanced and far ahead of what I know other websites use to catch cheaters." Hikaru Nakamura [and plenty of others]

As far as I understand it, recognising cheating is a matter of repetition and comparison, so highly unlikely to appear on just one move. It becomes a habit like anything else.

You would need to strike a deal with chess.com to see if it [or something similar] could be implemented on ICCF. Worth a try surely?

Some of the ICCF tourneys have entry fees and prize money attached, so do you have any detection tools for players with fake identities that cannot be verified? A 'shill' who might present simply to donate points/titles to other [real] players? 

I understand that on such is currently under investigation from a recent BCCC, so you are probably exposed to another form of cheating in any case.

Do something constructive to preserve your hobby - after all you want to continue to enjoy it!

Moving away from engine assistance would represent a fundamental shift in ICCF, and I am not sure that it would be well received by the purists (i.e. those who believe that CC is all about "finding the truth in a position")... it seems to me that this kind of chess is already well catered for elsewhere, so I'm not sure what additional opportunities ICCF would have to add to what is already available for players.

Yes, of course, we always investigate cases of collusion, although I am not aware of any currently active cases.

Yes it would be a fundamental move, but one well worth investigating.

The issue with the lofty 'finding the truth in the position' attitude is that it's no longer 'you' doing it - it's just a function of how powerful your PC is, and how long you allow it to analyze a position. 

The great advantage for a proposed change of this magnitude is that ICCF starts from such a high base, in terms of player numbers and status. It can open a dialogue with the likes of chess.com to develop an algorithm suitable for CC chess and offer non-engine chess as an option initially. You don't even have to ban players completely for suspected engine use - simply ask them to say, play another move/sequence of moves outside the top 5 suggested by SF when they are suspected.

As far as false identities goes, there is a fellow called 'Rhys Jones' under investigation at a recent BCCC is there not? AFAIK, there is no apparatus within ICCF which can either pick up on false identities or investigate them. 

Given the strength of open-source engines they are easy to create and maintain with the bare minimum of investment of time/energy by the creator.

I would like to see chess.com's sensitivity and specificity data before supporting such a proposal.

Obviously I can't discuss individual cases here, but I can assure you that there are no current investigations of Welsh players by ICCF.

Define "false identity" - do you mean someone playing under an assumed name? (if so, holding chess.com up as a model is amusing... there are plenty of woodpusher007's there ☺️)... or do you mean something else?

I have been reliably informed that there is such a case, and GM John Brookes is investigating it.

As far as chess.com's algorithm goes, you would need to mount a strong case to persuade them to reveal their I.P - they just wouldn't hand it over casually to any old joe. But you can start with lichess whose anti-cheat code is visible. Dedicated cheat detection neural networks may well materialize in the near future.

False identities are the same as multiple identities - players playing under a number of different personas with the ultimate aim of creating 'results' or false positives leading to titles and/or championships. There should be a detection tool for those in ICCF but there isn't one.

I have compiled a file of results which might point to suspicious activity.

P.S. I see David Philip Morgan has just 'earned' a CCM title norm by drawing all of his games in one tournament. Good luck to Phil but it is a great illustration of just how meaningless the whole situation has become...

John Brookes doesn't have an official role within ICCF or WCCF, if he is investigating anything it's nothing to do with ICCF.

Actually there is such a tool, we forensically analyse the server logs when we see any suspicious activity.

If anyone was "investigating" collusion by a Welsh player, they would have contacted both the ICCF Services Director for the forensic IT evidence, and the Welsh National Delegate for background details... since I hold both those positions, and haven't been asked about any cases in either position, I'd suggest that your source isn't quite as reliable as you think :-)

False identities are the same as multiple identities - players playing under a number of different personas with the ultimate aim of creating 'results' or false positives leading to titles and/or championships. There should be a detection tool for those in ICCF but there isn't one.

Nearly all WCCF and ICCF events are paid events - usually by PayPal... creating a PayPal account under someone else's name is almost impossible.

 A load of tosh Austin. Many people have multiple or secondary paypal accounts and payments are also invited via bank transfer and c/c in any case.

As for the rest of it, I'll continue compiling my file whenever I have the time and update it when other suspicious 'giveaway points' are flagged. I think I know who to send it too when it's ready.

The idea that blunders of the magnitude I have observed are accidental in this form of chess is very highly unlikely, and I doubt the apparatus or will is in place to monitor them despite your claims. As Paul said earlier they would probably be just written off as 'blunders'.

But hey, Happy Christmas in any case, it's just a hobby eh? :D

PayPal are obliged under anti-money laundering legislation to check the identity of the account holder, same with any bank or CC payment.

You can of course hold multiple accounts, but they must be in your own name.

Our own forensic procedures are also quite sophisticated... But, you asked what we had in place and I've told you. Not sure how any of that can be improved, but if you have any ideas of course I would be glad to hear them.

What I won't stand is witch hunting, bullying and persistent unsubstantiated accusations, something I have seen quite a lot of recently. That is totally unacceptable and the perpetrators of this harassment should be aware of the ICCF code of conduct and the consequences of their actions.

Merry Christmas,

Austin 

Not to mention the hurt and distress caused to those accused... We have lost at least one good official and several players due to to actions of these mindless thugs.