Mathematically impossible tournament results

[Huibertus]

I've had my doubts about the MP's tournaments result BBO calculates But now I have prove they must be wrong (at least sometimes).

 

Here are the hands.

 

 

https://webutil.bridgebase.com/v2/daylong_hands.php?tourney=ARDARD%3Ad3c18104.046c.11eb.b96d.0cc47a39aeb4-1601615102-&username=Huibertus

 

 

I had the feeling I did well, so I checked the results. Had the (shared) top result on 7 of the 8 boards, and the second best result (shared) on the 8th board was only topped by 1 player.

 

That means mathematicaly in MP's it is impossible to have finished any lower then (shared) runner up. And that only if the one player with the better score on the 8th board had the same top results on all other board, otherwise I must have finished (shared) winner.

 

 

However, according to BBO I finished 392nd !

 

Can anybody explain what the issue is, and how it is being adressed?

[pilowsky]

I'm sorry to tell you this, but what you have discovered is that you are a victim of BBO's crafty anti-cheating measure where multiple boards are used in these big tournaments.

Most of your apparently stunning results appear to have been in hands that everyone did well in - ?push boards? meaning that there is a ceiling on the percentage that your cohort can obtain.

On one board you "under-excelled".

 

It's devastating I know, but on the plus side it was free, you didn't have to pay and it cost you nothing. Already three good things!

Also, they did the bidding for you, so that's quite a saving in mental energy too. Seems like a bargain to me.

 

Imagine if it had been a "classic format". Then the robots would have played two of the hands as well.

[smerriman]

To avoid cheating, different players get dealt different hands (each hand is dealt a limited number of times so you're only compared to the small set who played it). There is nothing you can do about this; if you get dealt flat hands, you will finish well down the leaderboard no matter if you achieved the top score on every hand.

[pilowsky]

Like I said, which is why I prefer the challenge format.

Both the 'Just Declare' and 'Classic' formats seem to remove too much of the 'game' from the game.

To then aggregate everything into tiny groups just makes it a lottery.

Good players will consistently come out on top in the long run, but the randomness injected by the format is too great.

 

This why I prefer the Challenge format.

 

I don't know if it's possible to implement, but how about setting the daylongs so that each person has 5-10 minutes to play each board, but they can start the board whenever they want in the 24-hour cycle.

Cheating seems to be a minimal problem in the Challenge formats that run at the moment.

 

Algorithms exist to detect people that use the exact same line of play - real cheating is very hard. What would it look like?

 

Does BBO have any evidence that the implementation of this approach successfully prevented any cheating at all?

[barmar]

It's a "tournament" in the same way that downhill ski racing is a tournament. The snow conditions won't be the same from one run to the next, yet the competitors are still compared to each other.

 

And while you're not all being compared on the same boards, on each board you're being compared to a significant number of other competitors. Your final score represents whether you were playing overall well or poorly that session.

 

It's true that in short IMP games, one or two swingy boards can completely dominate the results; if you're not lucky enough to get any boards like that you'll have a hard time winning (but also a hard time coming in last if you find yourself on the wrong side of the swings). This is not as severe in matchpoints, since all boards have equal weight.

 

But the more boards you play, the more these random factors even out, as everyone should get about the same proportion of flat and swingy boards. The NABC Online is 72 boards, and at matchpoints this luck factor should be minimal (I won't go as far as saying negligible, but it could be). The fact that we consistently see the same collection of players near the tops of the leaderboards supports the hypothesis that skill is the dominating factor.

[thepossum]

I always assumed the large daylong format was less about cheating and more about giving everyone (EDIT correction - everyone capable of occasionally putting a few good hands together in a row) a chance of winning one

[mycroft]

I have three accounts. I mean, even if they didn't have different roles, I could have three accounts. I could have five.

 

I could run the daylong with one of my scratch accounts, then do it as "me" later. To catch IP tracing, I'll do it from different computers, and all the scratch accounts will only ever log in from McD's wireless.

 

I'm sure it *has* been done, which is why things are done this way.

 

Given the distressing level of "almost, but not quite, good enough to win" players we've seen getting "that little extra" now that sponsors and pros care about online online games are getting more attention, I think pre-emptively ensuring obvious tactics are covered is a necessary action.

 

Frankly, I have listened to people say that "20-30% of club players cheat" (give and read body language, listen to other tables, coffeehouse,...) for decades (along with the somewhat patronising "but as long as you don't expect bridge, it can be fun, and you'll beat them anyway." I've always thought that overblown (at least in its extreme form presented - definitely "getting that little extra" is there in measurable quantities). It looks like, either because when everyone starts playing screen regulations, or they have to play outside their little pond, or whatever, said players aren't quite as good as they think they are, a lot of them can't face that and find ways to get back to "their skill level".

 

It might not be 20-30%, but it sure looks like it's 20-30 in a thousand.

[pescetom]

It might not be 20-30%, but it sure looks like it's 20-30 in a thousand.

 

A few years ago, unexpected analysis of urine samples during medical certification of amateur sportsmen in Italy showed 6% positive for doping. In decades of competition and coaching in physical sports I came to conclude that 6-8% would cheat regularly, more or less independent of level or risk of discovery. Nothing I have seen so far suggests bridge is much different.

[barmar]

I suspect the 20-30% number isn't really about deliberate, collusive cheating, but also includes players who take advantage of UI without realizing that it's against the rules, perhaps subconsciously without even realizing it.

[mycroft]

Sure, and none of my examples were "collusive cheating" (some were deliberate, but I doubt many were deliberate and knowing). But it was interesting that the player that always sat at table 2, behind the resulters at table 3, had a habit of preempting on insanely short suits (4♥ on AQTxxx, 3♠ on AKxxx) *and always being right*. It was interesting that when North at table 1 made a "to play" bid, she flipped her card over and wrote in the contract, before partner's call. Never mind the "obvious signals", whether it was the extra push of the card, or the stare into partner's eyes to ensure she read it; the pairs playing a complicated system that always seemed to do better if the opponents asked about it; ...

 

And, as I've been saying for decades, this is an education problem, not a players problem. How to educate without beating on the novices; how to educate the TDs about both the problems and how to resolve them properly without causing offense; how to make people interested in "all that confusing law stuff"; those are the questions.

 

But as I said, it is interesting watching what happens when everyone suddenly has to play "screen rules". And what the crutch-users do when their crutches go away.

[pescetom]

But as I said, it is interesting watching what happens when everyone suddenly has to play "screen rules". And what the crutch-users do when their crutches go away.

 

When I moved our club onto BBO during the first wave of the pandemic, for most it was the first time ever without face to face crutches. One pair candidly admitted that they had had to relearn their bidding system from scratch because "they didn't have anything else to go on".