Latest Updates:
Page Index Toggle Pages: 1 ... 3 4 [5] 
Topic Tools
Very Hot Topic (More than 25 Replies) Dvorkovich on cheating (Read 26763 times)
Confused_by_Theory
God Member
*****
Offline


I Love ChessPublishing!

Posts: 662
Location: Europe
Joined: 05/13/15
Gender: Male
Re: Dvorkovich on cheating
Reply #3 - 08/25/20 at 16:47:22
Post Tools
Hi.

Mentioning DNA test level of certainty in many cases and near 100% certainty is going to frame the question somewhat for sure. Asked with less information included in the question answers could be slightly different.

But anyway. I find it OK to ask this way as long as Fide truly are in that ballpark. The problem is if this is not the case and only Fide can make that assessment atm.

Have a nice day.
  
Back to top
 
IP Logged
 
an ordinary chessplayer
God Member
*****
Offline


I used to be not bad.

Posts: 1807
Location: Columbus, OH (USA)
Joined: 01/02/15
Re: Dvorkovich on cheating
Reply #2 - 08/25/20 at 15:27:46
Post Tools
Confused_by_Theory wrote on 08/25/20 at 07:39:00:
1. Our methods of detection, although very advanced and ever-improving, can't provide a 100% confirmation. In many cases, the probability estimated is higher than the one for DNA tests. Do you believe a statistical algorithm (or a combination of those) giving close to 100% probability of cheating could stand as sufficient grounds for banning a player? If yes - what odds would you find sufficient?
In principle I believe banning should be allowed to happen based on strong statistical proof.
I would imagine at least some knowledge of how the anti cheating software puts together the near 100% probability is close to needed before giving opinions though. If say this remains near totally unclear for me as an outside observer I personally would be very much less inclined to a) run or ask for things to be run through the software in the first place and b) act on results.
Having multiple different statistical tools, as mentioned, is a good idea.

An algorithm needs to be verifiable. Researchers have made mistakes with algorithms before, sometimes quite crude mistakes, and so these things need to be checked by other researchers. This raises the issue that a cheater can inspect the algorithm and find a way not to exceed the threshold. So be it. If the algorithm is not sufficiently robust to survive such inspection by the enemy, then it can't be the sole basis for convicting such enemy. If an algorithm is a black box which can't be inspected or challenged, then an eventual miscarriage of justice is even more certain than the statistical "reliability" of the algorithm.

https://lostontime.blogspot.com/2019/10/ever-get-feeling-youve-been-cheated.html

People for some psychological reason find computational answers too convincing. Give someone a calculator, an algorithm, and a task to calculate the distance from the Earth to the Moon. If the calculator returned 1.496E+11 meters, too many people would without hesitation put that answer down. More decimal places is even more "convincing". Ask someone "how did you get that result?". If the answer is "we used an algorithm" (or "statistical model", or "computer program", etc.), too many people would be completely satisfied with the response and never check the result. I find the argument "higher than the one for DNA tests" to be basically an appeal for the audience to make this kind of acceptance error. The more I think about "DNA tests", the more I wonder how these can have any relation to algorithms for cheating.
  
Back to top
 
IP Logged
 
Confused_by_Theory
God Member
*****
Offline


I Love ChessPublishing!

Posts: 662
Location: Europe
Joined: 05/13/15
Gender: Male
Re: Dvorkovich on cheating
Reply #1 - 08/25/20 at 07:39:00
Post Tools
I'll give some opinions as well:

1. Our methods of detection, although very advanced and ever-improving, can't provide a 100% confirmation. In many cases, the probability estimated is higher than the one for DNA tests. Do you believe a statistical algorithm (or a combination of those) giving close to 100% probability of cheating could stand as sufficient grounds for banning a player? If yes - what odds would you find sufficient?
In principle I believe banning should be allowed to happen based on strong statistical proof. 
I would imagine at least some knowledge of how the anti cheating software puts together the near 100% probability is close to needed before giving opinions though. If say this remains near totally unclear for me as an outside observer I personally would be very much less inclined to a) run or ask for things to be run through the software in the first place and b) act on results.
Having multiple different statistical tools, as mentioned, is a good idea.

2. Shall FIDE apply sanctions for alleged online violations to over-the-board-play (and vice versa)?
Yes on both counts. To some extent when it comes to online though you need to be able to connect the player to the cheating account.

3. Shall we apply sanctions for alleged violations at platforms’ own events, and other unofficial online events, to official FIDE online events (and vice versa)?
I'd say ask the major online platforms and hear how they feel about this. To me it is close to a practical matter. Platforms should maybe follow some Fide standard for what is a violation etc. if this is the case though.

4. Shall we publish the names of alleged violators after the very first conviction?
I mean yea. It will have such an effect I feel so probably do it. Not for U18 though as they are literally young and dumb sometimes.

5. Shall the violators be punished retroactively, with their prize money, rating and titles been revoked for some period preceding the verdict? And, if yes, how far back should these actions go?
Yes. 1-2 years in normal cases more if there is proof of cheating before that as well.

6. What would you consider a reasonable banning period for first-time violators, and for repeat offenders? How strict should be the measures in youth competitions?
I think 2 years is ok for first timers. In principle I think lifetime ban for second. Cheating by U18 no sanction the first time, then 2 year ban, then lifetime. Exclusion if event where cheating took place is ongoing.

Have a nice day.
  
Back to top
 
IP Logged
 
Confused_by_Theory
God Member
*****
Offline


I Love ChessPublishing!

Posts: 662
Location: Europe
Joined: 05/13/15
Gender: Male
Dvorkovich on cheating
08/25/20 at 07:05:41
Post Tools
Hi.

Fide president Dvorkovich has published some questions about computer-assisted cheating in chess for the whole wider chess community.
https://www.fide.com/news/703

Somewhat interesting reading imo. Fide has faith in algorithms and is projecting complete readiness for legal challenges (defamation? retrieving prize money? something else?). I guess the main thing is what Fide does after this questionnaire. the questions asked are of course basis for probably some discussion though so let's go.

Questions from Fide:
1. Our methods of detection, although very advanced and ever-improving, can't provide a 100% confirmation. In many cases, the probability estimated is higher than the one for DNA tests. Do you believe a statistical algorithm (or a combination of those) giving close to 100% probability of cheating could stand as sufficient grounds for banning a player? If yes - what odds would you find sufficient?


2. Shall FIDE apply sanctions for alleged online violations to over-the-board-play (and vice versa)?


3. Shall we apply sanctions for alleged violations at platforms’ own events, and other unofficial online events, to official FIDE online events (and vice versa)?


4. Shall we publish the names of alleged violators after the very first conviction?


5. Shall the violators be punished retroactively, with their prize money, rating and titles been revoked for some period preceding the verdict? And, if yes, how far back should these actions go?


6. What would you consider a reasonable banning period for first-time violators, and for repeat offenders? How strict should be the measures in youth competitions?
  
Back to top
 
IP Logged
 
Page Index Toggle Pages: 1 ... 3 4 [5] 
Topic Tools
Bookmarks: del.icio.us Digg Facebook Google Google+ Linked in reddit StumbleUpon Twitter Yahoo