Measure claiming as proxy for goodness of play

[0 carbon]

One mark of a good player is that she knows when a contract is solid against any distribution partway thru the play and claims. So, the % of good claims is a measure of how good the player is.

I suggest BBO should record the number of good claims and the number of bad claims as a % of boards played and display it as

Claims: 4.67%/0.01%

on the player's profile page.

 

As well as showing how good a player is, this will also encourage players to claim earlier, speeding up play & tourneys.

A "good" claim is one which accords with GIB double dummy solver, whether or not it is accepted by opp(s).http://www.bridgebase.com/forums/public/style_emoticons/default/biggrin.gif A "bad" claim is one which does not.http://www.bridgebase.com/forums/public/style_emoticons/default/ph34r.gif

[Vampyr]

One mark of a good player is that she knows when a contract is solid against any distribution partway thru the play and claims. So, the % of good claims is a measure of how good the player is.

 

These forums might give the impression that bad claims are frequent, but in fact they are not, and most people will have none.

 

As well as showing how good a player is, this will also encourage players to claim earlier, speeding up play & tourneys.

A "good" claim is one which accords with GIB double dummy solver, whether or not it is accepted by opp(s)

 

This will in fact waste time, because if people want to boost their "good claim" rate they will claim on a two-way finesse or when there is a choice of squeezes etc, because GIB will get these right. But they will not be accepted by opponents, and might not be guessed correctly by the player when he has to play on.

 

And if defenders claim when they know that they have one trick, but declarer does not know that yet, do you still think that this is the mark of a good player?

[Kaitlyn S]

I voted no for a different reason. If people think that claiming is a mark of a good player, they will frequently claim or concede at trick 13 to run up their claim record.

[johnu]

I voted no for a different reason. If people think that claiming is a mark of a good player, they will frequently claim or concede at trick 13 to run up their claim record.

 

Well, at trick 13, one side can concede and the other claim on every hand, so the claim information is useless.

[helene_t]

hi 0 carbon,

 

measuring a players strength can probably be done more accurately by looking at their results in daylong tourneys.

 

but it might be good to warn opponents against a bad claim. there can be situations in which a "bad" claim should be accepted, for example when the director has decided the number of tricks or when a conditional claim was made. but mostly a "bad" claim should be rejected.

[The_Badger]

hi 0 carbon,

 

The true indication of player quality is how well they play with me, not whether they make correct or incorrect claims, I feel.

 

I would include in player quality too: politeness and willingness to stay at the table more than one board, and leaving in the middle of boards, etc.

 

Most decent, in manners and standard, players on here will make correct claims. And we all make mistakes occasionally.

 

What annoys me more is when opponents - especially those labelled intermediate and above - refuse, sometimes repeatedly, genuine claims. Maybe we should perhaps have a 'stats for prats' figure on their profile instead.

[aawk]

I do not think showing the % of a players claiming history is a indication of their level of play for a few reasons.

 

It does not show the moment the claim is made. Meaning anyone can claim trick 13 when they could have claimed at a earlier stage.

 

What is a wrong claim ? Is it if opponents reject a claim then there will be players who will reject every claim to mess up a players claim %. Or what will happen when opponents accept a wrong claim.

 

Who will decide if a claim is good ? At live play a director must be called when a claim is rejected and the players cannot get to a agreement and according to the rules you are not allowed to play on after a claim. Playing online you would need a computer program to be a director (would be great if this was possible though).

 

Is it a relevant statistic ? I think it is but not as a indication of a players playing level.

 

Showing a players scoring avarage (MP, IMP, team results) or some other form of rating would be better.

[helene_t]

I do not think showing the % of a players claiming history is a indication of their level of play for a few reasons.

 

It does not show the moment the claim is made. Meaning anyone can claim trick 13 when they could have claimed at a earlier stage.

 

What is a wrong claim ? Is it if opponents reject a claim then there will be players who will reject every claim to mess up a players claim %. Or what will happen when opponents accept a wrong claim.

 

Who will decide if a claim is good ? At live play a director must be called when a claim is rejected and the players cannot get to a agreement and according to the rules you are not allowed to play on after a claim. Playing online you would need a computer program to be a director (would be great if this was possible though).

To be fair, BBO already has an algorithm that validates claims (the one GIB uses to accept or reject claims). There can be situations in which it misclassifies a claim (for example, if you have a feeling that one of two high cards are good but you don't remember which one, you can claim one trick - the software will assume that you remember correctly). But largely the algorithm works.

 

One could obviously require the claim to be stated at (say) trick 10 or earlier.

 

Whether opps accept or reject should be irrelevant.

 

That said, as I already stated, I don't think the "correct claim" % would be a very useful statistic.

[manudude03]

I suppose it could work with some modifications. Trick 13 claims do not count. It is counted as a valid claim if there is a line that guarantees you will take as many tricks as you claim.

[SelfGovern]

Two things:

1) Claims on BBO are already a mess sometimes. I can't count the number of times I've seen a concession of all the tricks rejected (either by defense or declarer). I know I have more than once personally tried to concede all the tricks two or more times, only to have it rejected. So then I concede all but one... and it's often accepted. So odd.

 

2) Your "no" vote says: "No, I don't want this measure made of me" -- I don't mind people measuring me against this; I try to claim when it's clear and state a line of play... but I don't think it's particularly useful for the majority of people. Other valid reasons for a "no" vote would be:

- you didn't want people gaming the claim system (which could serve to slow play in the end),

- you don't think it's a useful measure of skill

- it's often not useful unless declarer states a line of play and GIB uses this in the evaluation of valid claimhood

 

 

But it is an interesting question.

[Caitlynne]

You measurement argument does not pass the sniff test. There are just too many confounding variables that introduce error and invalidate your proposed measure.

 

Two (among many) counterargument examples:

 

1. Your %good measure does not account for how deep into the hand a claim is made. That surely matters in terms of measuring skill.

 

2. Perhaps I see the claim quite early and yet, fearing the opponents would either not see it or, if they did, would not trust that I see it, they might refuse. Therefore, perhaps I have a disincentive to claim early. Which means that your measure is flawed.

 

Like the Star system, your proposed measure is a poor indictor of skill.

[msjennifer]

The idea is certainly good.But many details will have to be worked out before it can satisfy everyone.Suggestions ,as to what points are to be worked out ,may be invited from all and then an expert advisory can formulate a scheme agreeable to all.

[johnu]

It is counted as a valid claim if there is a line that guarantees you will take as many tricks as you claim.

 

A claim that requires double dummy play shouldn't be counted. While a player could have every card placed correctly, the hand may still be up in the air in real life. e.g. There could be a 2 way finesse that is never misguessed double dummy, or something exotic like a compound squeeze that depends on guessing the opponents distribution that could easily be misplayed. Since GIB doesn't do single dummy analysis, that puts into question every claim analysis by GIB.

[aawk]

To be fair, BBO already has an algorithm that validates claims (the one GIB uses to accept or reject claims). There can be situations in which it misclassifies a claim (for example, if you have a feeling that one of two high cards are good but you don't remember which one, you can claim one trick - the software will assume that you remember correctly). But largely the algorithm works.

 

One could obviously require the claim to be stated at (say) trick 10 or earlier.

 

Whether opps accept or reject should be irrelevant.

 

That said, as I already stated, I don't think the "correct claim" % would be a very useful statistic.

 

 

If the algorithm works in most cases then it is good enough for me. If the claiming % should be in your profile is the remaining question. I would not mind aslong there is a option for any player to hide it if they want to.

[Kaitlyn S]

I have made several bad claims on purpose to save time.

 

I'm in 4S on a heart lead encouraged. I know the opponents are taking 2 hearts, then I'm drawing trump and taking the diamond finesse. I will claim 5 immediately and say Please reject if East has the DK - I'm taking the finesse.

 

If it's rejected, I will actually draw trump and take the finesse since once in a while it will get rejected even when West holds the DK either because they don't see it or because they are hoping I'll claim 4.

 

Since I'm doing most of my play in team matches nowdays, half of these "conditional" claims are getting accepted and on we go to the next hand instead of that boring one we were on.

 

You might say that it might save time just to draw the trump. That would be true if nobody had to discard; but someone might take some time to ponder their discard(s) when it doesn't matter, why make him do that? This way, even if the finesse is off, he can see my hand and can discard quickly.

[1eyedjack]

It might work if people didn't know you were doing it.

[ShirleyMqz]

measuring a players strength can probably be done more accurately by looking at their results in daylong tourneys.

 

Percentage in the daylong tourneys isn't accurate either. Some players play them attempting to maximize their probability of hitting a high spot on the leader board rather than maximizing their average result. That means things like bidding risky slams, and playing for risky overtricks in matchpoint games. If those ploys work it improves your chances of hitting the top 10 or top 100 or whatever; if it fails it just means you're mired back in the pack like you were going to be anyway.

 

Those daylong tourneys aren't anything like a fair test of skill in any case because not all the players get the same hands. If you don't get the slam hands you're never going to hit the top of the IMP or total point tourneys no matter how well you play. If you don't get the hands with a lot of potential for matchpoint swings you won't win the MP tourneys.

[cynac]

I agree with poster who said no, not because he did not wish this measurement but for other reasons. There is almost no merit in the notion that making good claims marks out a good player. And if your algorithm really does say that if there is a winning line then that is a good claim, then that is clear nonsense. Almost every bridge-column-problem has a winning line. Would put them out of business if the answer to every problem was "claim" - and the computer will say that's right.

[GrahamJson]

I'm not convinced that claiming saves much time, unless there are several tricks remaining and the claim is very obvious. Often it is quicker to play out the remaining two or three tricks quickly rather than do through the motions of claiming (at least it can be if using an iPad).

 

Claiming on BBO is not like in real life bridge, when you can explain your line of play; "draw trumps and dummy's clubs are good". I've seen claims when I am not at all sure that declarer knows what's going on, and it's only when the claim is rejected that he realises that there is a trump out, for example. I've also seen a defender concede the remaining tricks, quickly accepted by declarer, when in fact he would have taken none of them. (In one case I was the defender's partner, holding the remaining trump, and the concession was accepted before I could reject it.)

 

I think that, unless they have changed, the rules state that if you don't make your line of play clear you cannot draw trumps to take a finesse after a claim. If you have to type in an explanation to cover this when making a claim it would slow things down, not speed up.

 

If you want a player's statistics shown on his profile how about; % of hands left before play complete; average number of boards played at a table before leaving; average imps per board; average number of boards played by their partners before leaving; average number of boards that their oppo stay before leaving. I believe that the better players (at least better behaved, if not the best technically) will tend to stay around longer with partners and oppo who are happy to play with them.

 

One last thing. If a player leaves mid board and the deal is subsequently redealt I suggest it should be completed by GIB in the background and any negative score allocated to the leaving player. I have seen too often players make bad bids or plays and then immediately leave when it is clear that they are heading for a bad score.

[barmar]

Since GIB doesn't do single dummy analysis, that puts into question every claim analysis by GIB.

GIB uses single dummy analysis when deciding whether a claim is good. I don't think it ever accepts a claim based on guessing a 2-way finesse correctly.

[Vampyr]

GIB uses single dummy analysis when deciding whether a claim is good. I don't think it ever accepts a claim based on guessing a 2-way finesse correctly.

 

Its single-dummy analysis is pretty poor too. I recently played a little with the robots, and often when I claimed with all top tricks the only option I was given was to concede the rest of the tricks.

 

This type of thing would make the OP's suggestion spectacularly unhelpful.

[helene_t]

Its single-dummy analysis is pretty poor too. I recently played a little with the robots, and often when I claimed with all top tricks the only option I was given was to concede the rest of the tricks.

I believe its algorithm is very good. There is a flaw in the implementation which allows only concedes when opps lead to the trick but that's a different issue.

[Kaitlyn S]

This whole discussion brings up an interesting point: the OP wants a percentage of claims to see how good the player is?

 

1. There are players who have played for many years that might be able to claim accurately but they are wretched bidders, so bad that there is little correlation to what they are bidding and the cards they hold.

 

2. Let's say that you could determine the skill level of the player. That's not all that should matter to your enjoyment. I have played with many players here that are good but are not at all fun to play with - they are very critical about many bids and plays and not always right despite their skill, and while they are taking time to explain what you did "wrong", the opponents are griping about your partner slowing down the game (and you too if you respond.)

 

Not all, but most of the players I have found enjoyable to play with here are weaker players than I am (I think.) I might do less well in the IMP score than I would with an arrogant partner that deserves his "expert" rating, but we might play somewhat the same system, and any discussion is constructive (usually either he or I will way "where could I have improved the auction"?)

 

Quite frankly, a partner who interprets bids and defensive carding the same way I do but fails to make a hand on a baby squeeze that I would find in my sleep is close to an ideal partner, because the parts where we are communicating go swimmingly.

 

Do the players who want to know others' skill levels want to find weak opponents so they can run up their IMP score and look better than they are? To me, that's not much fun. The games I enjoy the most are where the opponents are decent.

 

I can see someone using skill level as a starting point to exclude certain players but in the long run, if you want to find partners that you really mesh well with, you're going to have to keep trying people until you find some you like. That's what I'm doing, and it's an interesting road. However, I am currently following some players marked as "Intermediate" because I enjoy playing with them.

[GrahamJson]

As far as the self assessed skill level is concerned, it is pretty useless. I don't think that I have seen a single "expert" who is even close to being that. Most would struggle to get 50% in any decent club. I previously a few suggestions for statistics that might be useful. None are an indication of expertise, however they might indicate how enjoyable a player might be as a partner or opponent.

[Kaitlyn S]

As far as the self assessed skill level is concerned, it is pretty useless. I don't think that I have seen a single "expert" who is even close to being that. Most would struggle to get 50% in any decent club. I previously a few suggestions for statistics that might be useful. None are an indication of expertise, however they might indicate how enjoyable a player might be as a partner or opponent.

There is one useful statistic that is already recorded that may give an idea of enjoyment but I can't see it either because (a) I am too new, (b) I can't use the client, or © I'm ignorant of ow to see the information.

 

That factor is board completion rate. I can see my own - it is 99% (must be rounded down as I have played over 200 hands and left once) and the one time is a time when my partner on defense asked for an undo 25 times and was refused 25 times and left, and I didn't relish the opportunity to finish defending the hand with someone who knew nothing about the early defense (I asked for a redeal several times and was likewise refused; we were headed for a wonderful score if my partner remained someone who knew the deal but were probably getting an awful score with someone who knew nothing about the hand - something the refusing player probably knew and was just being despicable when not redealing. Essentially I left because I thought his actions were almost cheating.)

 

In any event, I don't know if I can see the completion rate for others, if someone completes 60% of their deals, it means they are leaving at the first sign of trouble, like if you misguess a finesse, or even if an opponent annoys them, leaving you again to defend the last half of a hand with someone that doesn't know the early defense. Several pieces of **** avoid reducing their board completion rate by always hosting and booting their partners that misguess a finesse or their opponent that takes more than a second to play a card, so even this statistic is dubious, unless it's combined with a statistic about booting other players from his table.

 

There is other questionably useful information that others can see but I haven't figured out how to see it. Apparently others can see my IMPs per hand for both the last 50 deals and for the last month. I have not figured out how to see that yet, and wouldn't be surprised if I can't because the client is not available to me. However, I am sure others can see it because someone posted that I had an impressive IMP ratio. TBH it was greatly inflated because I was trying everything including playing a lot in the relaxed room, where the opponents were twice as likely to have a 14-IMP disaster than partner. Full disclosure: I'm not an awesome player :)

 

Even if I got that information, I don't think it would be that useful. For example, I can run my IMP ratio up by playing in the relaxed room a lot. Now I'm playing mostly team matches; the competition is a little better so my scores are a lot less. I also am picking up partners there which hurts my score. If I play a lot with the same person and we make agreements, my scores will increase again. I'm the same player in every situation but my IMP ratios are all over the place. I imagine the same is true for everybody else, not to mention that players who play against strong opponents a lot are going to appear to be a lot less skillful than they are.