Success percentage

[HighLow21]

What is a good percentage of declared contracts to make? I'm aware that if you make 100% of your contracts that means you're not bdding enough; what's a reasonable rule of thumb for good play? 60%? 2/3? 75%?

[Fluffy]

will depend on how much you like to save and the level of your opponents as well as their willing to compete further.

 

I think you should rather compare how many contracts you make as opposed to how many opponents make.

[Lorne50]

There is no one answer.

 

1. The odds are different for IMP scoring vs MPs

2. You have to distinguish between contracts you bid to make with competitive auctions where going off may be a great score

3. Even if it is an uncompetitive auction it may not be clear. Suppose your outstanding relay bidding system tells you 7♠ is a 60% contract with your combined 24 count. If only half the field bid 6♠ with the rest in game you would be silly to bid it since you lose 75% of the available MP's 40% of the time but only gain 25% for the 60% of the time it makes.

[billw55]

I think Pavlicek has some stats on this. As I recall, most of the top pairs were between 60-70% making.

[HighLow21]

I think Pavlicek has some stats on this. As I recall, most of the top pairs were between 60-70% making.

Will have to check Pavlicek. Thank you!

[HighLow21]

There is no one answer.

 

1. The odds are different for IMP scoring vs MPs

2. You have to distinguish between contracts you bid to make with competitive auctions where going off may be a great score

3. Even if it is an uncompetitive auction it may not be clear. Suppose your outstanding relay bidding system tells you 7♠ is a 60% contract with your combined 24 count. If only half the field bid 6♠ with the rest in game you would be silly to bid it since you lose 75% of the available MP's 40% of the time but only gain 25% for the 60% of the time it makes.

I guess what I'm looking for is just a blended, long-term average. A range. I recognize IMPs will probably be a bit lower if played correctly (more sacrificing and 50/50 game bidding) and I also recognize that pairs that sacrifice more (even when it is best to do so) will fare worse in this metric. It's part of a multi-factor model I'm looking at:

 

1. How often do you get to the best level (part-score, game, ss, gs) relative to your opponents

2. How often do you make that contract relative to your opponents

3. How frequently do your doubles and redoubles work out

 

etc.

 

This kind of review helps me gauge my overall improvement and decision making.

[HighLow21]

This is extremely relevant and interesting, to billw55's point. Thank you!

 

http://www.rpbridge.net/9y18.htm

[HighLow21]

This is extremely relevant and interesting, to billw55's point. Thank you!

 

http://www.rpbridge.net/9y18.htm

Lots of other great stats too -- like average IMPs per board as declarer or on opening lead by player. Average across all the players was 65.88%, which lines up with my original idea that it should be around 2/3 of the time.

[Free]

There is no one answer.

 

1. The odds are different for IMP scoring vs MPs

2. You have to distinguish between contracts you bid to make with competitive auctions where going off may be a great score

3. Even if it is an uncompetitive auction it may not be clear. Suppose your outstanding relay bidding system tells you 7♠ is a 60% contract with your combined 24 count. If only half the field bid 6♠ with the rest in game you would be silly to bid it since you lose 75% of the available MP's 40% of the time but only gain 25% for the 60% of the time it makes.

Good response. One thing you've forgot to mention though: it's also different when you're NV or V (clear example is imp games).

[JLOGIC]

Can someone explain this to me?

 

http://www.rpbridge.net/9y16.htm

 

I make quite few contracts compared to the field in this, I am one of the lowest at ~59%, but my imps/bd and imps/par seem to be one of the highest. My first thought was maybe I play a lot of partscores down when they make partscores or something, but that would mean I declared a high percentage of hands and I only declare 26.29 which is above avg but not an insane amount.

 

How can this be? And sorry this is not meant to be some kind of brag post I genuinely do not understand how I can do well declaring while making far fewer contracts and declaring a relatively normal amount of hands.

[JLOGIC]

http://www.rpbridge.net/9y17.htm

 

This one is even worse, I make 53.2 %(!) of my hands and declare 26.38 %, but again have one of the best imps/bd and imps/par when I declare. I don't think this is a sample size thing, that's like 1000 total hands of me making very few hands. As an example, Fantoni with about the same % played and imps/bd imps/par made 72 % of his hands. Bocchi and Madala in the first sample made 68-69 % with similar numbers. Wtf?

[PhilKing]

Can someone explain this to me?

 

http://www.rpbridge.net/9y16.htm

 

I make quite few contracts compared to the field in this, I am one of the lowest at ~59%, but my imps/bd and imps/par seem to be one of the highest. My first thought was maybe I play a lot of partscores down when they make partscores or something, but that would mean I declared a high percentage of hands and I only declare 26.29 which is above avg but not an insane amount.

 

How can this be? And sorry this is not meant to be some kind of brag post I genuinely do not understand how I can do well declaring while making far fewer contracts and declaring a relatively normal amount of hands.

 

My guess is that imps to par as declarer are based on absolute par rather than the par for the contract you are in (which is why everyone is minus), so being in the right contract is the most important factor.

[JLOGIC]

How can one be in the right contract a lot if they are frequently down? I don't get it. Unless par is saving frequently or w/e in which case it would seem like I would play a high amount of hands (I'm pretty sure I don't save frequently).

 

Sorry for derail I just don't really get it.

[mikeh]

Can someone explain this to me?

 

http://www.rpbridge.net/9y16.htm

 

I make quite few contracts compared to the field in this, I am one of the lowest at ~59%, but my imps/bd and imps/par seem to be one of the highest. My first thought was maybe I play a lot of partscores down when they make partscores or something, but that would mean I declared a high percentage of hands and I only declare 26.29 which is above avg but not an insane amount.

 

How can this be? And sorry this is not meant to be some kind of brag post I genuinely do not understand how I can do well declaring while making far fewer contracts and declaring a relatively normal amount of hands.

Maybe it is because you bid more games and slams than most. As such you rate to be overbidding compared to the rest of the field and thus making fewer contracts, but every game you make that was missed at other tables pays a lot more imps than you lost by going down on the others. It's just a question of fine-tuning degrees of aggression. Too much and not only will you be failing more than most, but you'd be losing imps as well. Too little, and you make most of your contracts, but don't pick up many imps.

[quiddity]

Maybe your teammates defend more brilliantly when they suspect you'll be declaring. :-p

[HighLow21]

Maybe it is because you bid more games and slams than most. As such you rate to be overbidding compared to the rest of the field and thus making fewer contracts, but every game you make that was missed at other tables pays a lot more imps than you lost by going down on the others. It's just a question of fine-tuning degrees of aggression. Too much and not only will you be failing more than most, but you'd be losing imps as well. Too little, and you make most of your contracts, but don't pick up many imps.

Justin we've established in previous posts that you tend to be more aggressive in competing an auction--perhaps also in bidding borderline games and slams. What I suspect is that your bidding is more aggressive both for sacrifices and for makeable contracts that other tables don't find. For example, if you get to a 17-HCP spade game that makes 50% of the time and is a good save against their rock-solid 4♥, or something like that.

 

It's my suspicion that when one gets to the highest levels of the game there's not a significant relationship between percent of contracts made and IMPs earned per board. Perhaps even a negative relationship in certain situations. We're talking about a lumped average, of course, so there are many ways to pick apart the data and many different factors contributing to the results.

 

One clue to me is that your I:pad stat is surprisingly high (-1.27 vs. a field average of -1.99). My guess is this: you bid mre aggressively, and more frequently find good "saves" for your side; you frequently overbid, but that's balanced out by overbidding to MAKING contracts that are either fortunate in the lie of the cards or misdefended. This goes well with Bob Hamman's philosophy of bidding game to put pressure on the defense because defense is difficult.

 

Just my thoughts

[HighLow21]

Also Justin, if you draw a scatterplot of I:pad vs. % making for the results in this data set, you get virtually zero correlation:

 

http://www.rpbridge.net/9y16.htm

 

But check out Jonsson,T. 51.79% making, but -0.71 results vs. par. He only plays 26% of hands. An excellent sacrificer? Bids and makes a high number of low-HCP games the field doesn't find? I leave it to you.

[JLOGIC]

Also Justin, if you draw a scatterplot of I:pad vs. % making for the results in this data set, you get virtually zero correlation:

 

http://www.rpbridge.net/9y16.htm

 

But check out Jonsson,T. 51.79% making, but -0.71 results vs. par. He only plays 26% of hands. An excellent sacrificer? Bids and makes a high number of low-HCP games the field doesn't find? I leave it to you.

[/quote

 

That is almost identical with my sample with kevin in 2012 http://www.rpbridge.net/9y17.htm 26 % hands, -.79 and 53.23 % making. I don't understand it, sacrificing doesn't make that much sense since it seems like I would declare more frequently.

 

What you guys said about bidding more games/slams makes some sense, but it is hard for me to believe I bid that many more games than the people in this sample, everyone at the top level bids a lot of games. But that's the only thing that makes sense I guess...thanks. Interesting data for sure.

[mikeh]

Also Justin, if you draw a scatterplot of I:pad vs. % making for the results in this data set, you get virtually zero correlation:

 

http://www.rpbridge.net/9y16.htm

 

But check out Jonsson,T. 51.79% making, but -0.71 results vs. par. He only plays 26% of hands. An excellent sacrificer? Bids and makes a high number of low-HCP games the field doesn't find? I leave it to you.

 

That is almost identical with my sample with kevin in 2012 http://www.rpbridge.net/9y17.htm 26 % hands, -.79 and 53.23 % making. I don't understand it, sacrificing doesn't make that much sense since it seems like I would declare more frequently.

 

What you guys said about bidding more games/slams makes some sense, but it is hard for me to believe I bid that many more games than the people in this sample, everyone at the top level bids a lot of games. But that's the only thing that makes sense I guess...thanks. Interesting data for sure.

 

 

I think it important to bear another point in mind.

 

The statistics are of small samples: no partnership played even as many as 1000 hands and most played far fewer. The minimum requirement was 200 hands, and that means that for some players/partnerships, the declarer statistics are based on maybe 50-60 hands played. I do recognize that for Justin and others the number is closer to 200, but 200 hands represent but a tiny, tiny fraction of declarer play experience for these players, even if one limits the analysis to 'high-level' competition or practice.

 

I leave it to the more mathematically inclined to opine what a 'good' sample size would be before one could have, say, a 95% level of confidence that a particular metric allowed for a firm conclusion to be drawn.

 

Meanwhile, it remains plausible, to me if to no-one else, that in this small sample, one partnership might bid more games, on a relative basis, than most. As one example, maybe the preemptive hands suited some partnerships better than others, and over 300 or 500 or even 700 hands, there probably wouldn't be more than a handful of hands where that mattered.

 

I'm not sure that one can correlate, however, the bidding of a game as always or even mostly arising from the bidding tendencies of declarer, as opposed to the partnership.

 

Sure, opening 1♠ may lead to 4♠ (or to any spade contract) more often than would passing it, but as against that, often times the reason declarer is playing the hand is because partner transferred, or raised opener aggressively or made an aggressive takeout double....say all would open, but half the field invited and half forced to game. Declarer scores well when it makes and badly when it fails, but getting to or missing game wasn't his choice.

 

I guess what I am trying to say is that one can perhaps draw some inferences as to how various players and partnerships performed in those events in that year, but that the sample size seems too small to me to allow for much beyond that.

[JLOGIC]

True. If I wasn't lazy maybe I would try to keep records of this when I play, I doubt I will actually follow through with it but might be interesting to do it for a year and get thousands more hands of data. In poker you have a lot more hands to work with but you can often see leaks or strengths just based on looking at the data, bridge would probably be harder to do that with but it still might be useful.

[wank]

My guess: playing precision you're getting to game without revealing your hand. I don't know how much relay stuff you play but this may be benefical to your slam bidding. This must result in big gains on game hands. Precision handicaps your part score bidding though - less space to diagnose the right contract. Losses in part score are more numerous but gains from game are larger.

[HighLow21]

I think it important to bear another point in mind.

 

The statistics are of small samples...

Ya I was beginning to notice that the sample was small. For Justin and a number of others the number of hands was 300-400, so about 75-100 hands declared. If the standard deviation of IMPs/board is as low as 3, then the standard error will be around 0.3 or 0.4 or so. That means a fair amount of variation. In my own datasets, when I do any kind of breakdown I tend not to jump to conclusions unless I'm including at least 1,000 hands... and for some measures (like slam performance) I need a lot more because they happen relatively frequently.

 

Still some data is better than nothing, and if the trends hold, perhaps there's an explanation for it -- probably some combination of the things we've identified plus 1 or 2 others we haven't.

[JLOGIC]

Ya I was beginning to notice that the sample was small. For Justin and a number of others the number of hands was 300-400, so about 75-100 hands declared.

 

The 2 links I found had 235 and 704 so 939 hands. Point still remains though.

[HighLow21]

The 2 links I found had 235 and 704 so 939 hands. Point still remains though.

Now we're talking about close to 250 hands declared, so I think it's reasonable to say that there's something to this. Who were your teammates in 2011/12 out of curiosity?

 

It looks like it was Hurd and Woolridge. They were setting a lot of contracts at the other table (about 42.5% vs. 36.5% average for the field). This might help explain the strong I/bad statistic; I:pad should, I think, be independent of the other table. Anyway, it's reasonable to assume it had something to do with (1) finding more making contracts that the opponents your way did not find, and (2) finding more good sacrifices than the opponents your way.

[billw55]

Maybe it means that you are getting away with way too much. Maybe going down 50 a trick often, opponents aren't doubling enough, or selling out too easily. Note to self: be ready with chopper when Justin comes to table :P