Performance Index

[micsfyuen]

Today, I looked at the results of last months through Myhands. This time I sorted the scores differently to see if anything interesting comes up.

 

I copied the scores into an Excel spreadsheet, keeping only the boards that I partnered my favorite partner, and sorted the IMP scores into 4 columns according to who declare the hand. (I play, partner plays, I lead first, partner leads first) Then I calculated the average IMP/board of the 4 columns to see who the better declarer is/who consistently make losing first leads.

 

I was pleased to see that all 4 averages are positive, but to my surprise, these 4 scores differ by quite much. The boards that we declared scored much better than those we defended on average, and boards that I played scored quite differently as partner's.

 

I tried this analysis on 3 other regular partnerships that I know. Interestingly, all of us scored better in declaring than in defending. A friend of mine scored a stunning +3.13IMP/board average on the boards that he declared, but only +1.27 IMP/board on defending.

 

I am not a good statistician, so I may not be able to comment on how significant these differences on the scoring average are. But I believe this kind of analysis would bring us insight on how to improve our game and which skills to work on.

 

Myhands is such a wonderful tool for us to keep tracking our performance. May someone be able to make use of the data to come up with a performance index or the like to help everyone understands their strength and weakness?

[billw55]

It's a common effect at pairs, to get better results for declaring. Making a normal contract, is usually a good score, because there are always a few who botch it.

[jjbrr]

Also sort them by whether or not your score was plus or minus, and I believe you'll see that "a plus score is a good score" has quite a bit of truth in it.

[hotShot]

There are [switching to political correctness mode] impolite players who punish partners errors with useless redbles and overbids. As a result these pairs get a very bad score, pushing the results of the declarers up.

Since the lucky defenders of the deals get an undeserved top, that no one else can reach, it pushes the IMP score of defenders down, that reach a normal result.

[jjbrr]

There are [switching to political correctness mode] impolite players who punish partners errors with useless redbles and overbids. As a result these pairs get a very bad score, pushing the results of the declares up.

Since the lucky defenders of the deals get an undeserved top, that no one else can reach, it pushes the IMP score of defenders down, that reach a normal result.

I'm not convinced this happens nearly enough to be the main culprit. I think it's much more common that randoms just have misunderstandings and stop in partscores in grand hands.

 

Any making game will usually win imps for your side.

[helene_t]

It's a common effect at pairs, to get better results for declaring. Making a normal contract, is usually a good score, because there are always a few who botch it.

It could be that people tend to underbid so that deciding to defend is more likely to be wrong than deciding to declare. I find that plausible for matchpointed club games. I find it less plausible on BBO where people tend to bid too much and also it's IMPs so a too timid balancing strategy would be punished less severely.

 

Rather I think Jeremy has it right. Bad players can make more costly mistakes on the deals that belong to them, giving the better players an extra advantage when thet hold good hands.

[Finch]

For any hand that looks like a 'normal game' hand, look at all the 16 results and there will be some insane ones (either gross stupidity, or partnerships not on the same wavelength). As a result bidding & making game is virtually always worth somewhere between 3 and 8 imps, and defending the the opposite effect.

[billw55]

Bad players can make more costly mistakes on the deals that belong to them

I agree. But I don't see how that is much different from "there are always a few who botch it". :P

[Mbodell]

It's a common effect at pairs, to get better results for declaring. Making a normal contract, is usually a good score, because there are always a few who botch it.

I've been tracking all of my results for face-to-face matchpoint hands for the past three and a bit months and I've been surprised to see that this[declaring means good score] isn't true for me. I've got nearly 1500 hands that I've tracked (ranging across 4 regular partners and 7 occasion partners and with club, sectional, regional, and national events) and I'm around 0.5% better when our side is defending than when our side is declaring.

 

Also sort them by whether or not your score was plus or minus, and I believe you'll see that "a plus score is a good score" has quite a bit of truth in it.

 

This is certainly true. My breakdown of hands gives me 66.7% for positive scores and 36.5% for negative scores (and 58.4% for pass outs although I only have 5 of these so large margin of error here).

 

There are a couple of other things I've noticed that are a little odd.

 

One interesting oddity that is interesting is my 2 most common partners (with whom about 2/3 of my hands are with and both of whom I consider my peers in skill level) insist on sitting N and E so I most often sit S and W (and we are usually E/W because we are young people). I've noticed that the stronger player in a pair often sits N (and to a lesser degree E), especially at the club level where there are more widely unbalanced pairs. And this means that on these hands the better player of the opponents is on lead when I declare, and that when I'm on lead the worse of the opponents is declaring. I bring this up because there is a little over a 2% difference in my results where if my partner is declaring we score a little over 2% better than when I'm declaring. But when I'm leading we score a little over 2% better than when my partner is leading. It could be that my partners are better declarers than me and that I'm a better opening leader than they are, but it could also just be that we play enough unbalanced opponents that sit this way that it effects the results based on the opponent who is leading/declaring. I'm interested in tracking these trends as I get more hands.

 

A different thing I've noticed has to do with strain. As is somewhat expected when my side declares we declare ♣ < ♦ <<<< ♥ < ♠ << NT. And our sides matchpoint results are similarly ordered with better scores for NT, then spades, then hearts, then diamonds, then clubs and again there are noticeably bigger gaps between NT and the majors and between the majors and the minors. When I'm defending, on the other hand, my opponents most frequent strain is ♥, although it is more of a distribution of occasionally the minors and frequently any of the other three suits. But for my results I get the exact same scores for defending all of the opponents contracts, except ♥. Defending ♥ is noticeably worse for my side then defending any other strain.

[PhantomSac]

You get more MP on defense than declaring? That is truly bizarre and pretty hard to believe (not that I don't believe you, it is just so unusual that it shocks me).

 

Edit: I thought about this and came to the conclusion that you must compete way too much so that you are only defending when it's very clear to do so, and are thus hurting your average when declaring and boosting your average when defending. Any chance this is the case? What % of hands are you declaring? Obv there are other possible explanations like you bid very badly in general and are a good card player, so defense is good for you since your bidding doesn't matter whereas you usually just declare bad contracts, but I doubt that's the case obv.

[Mbodell]

You get more MP on defense than declaring? That is truly bizarre and pretty hard to believe (not that I don't believe you, it is just so unusual that it shocks me).

 

Edit: I thought about this and came to the conclusion that you must compete way too much so that you are only defending when it's very clear to do so, and are thus hurting your average when declaring and boosting your average when defending. Any chance this is the case? What % of hands are you declaring? Obv there are other possible explanations like you bid very badly in general and are a good card player, so defense is good for you since your bidding doesn't matter whereas you usually just declare bad contracts, but I doubt that's the case obv.

Well, a little over 40% of my hands are with my most common partner with whom I play an anti-field strong club mini/weak nt with so we may well be more active than most people and a hypothesis like that might be true.

 

However, looking at the data, I actually defend slightly more than I declare as I have ~51% of the hands (both with all partners and with the above partner) we defend and ~49% of the hands we declare.

 

Maybe I'll see it is a fluke and that a 0.5% advantage for defending over declaring in my first 1500 tracked boards will disappear as I get more data. But it was a counter intuitive result even to be so close, and something like the +ve versus -ve scores was so clear in comparison.

[hotShot]

66.7% for positive scores seem very good, but obviously you are bidding a little to aggressive at some points so that you have only 36.5% when the score is negative.

I read that that you are declaring and going down, when they are going down as well.

You should focus on that, you can gain a lot at this problem.

[Mbodell]

66.7% for positive scores seem very good, but obviously you are bidding a little to aggressive at some points so that you have only 36.5% when the score is negative.

I read that that you are declaring and going down, when they are going down as well.

You should focus on that, you can gain a lot at this problem.

Sorry, I think you are maybe reading that wrong. The 66.7% is for any positive score (our side making a contract OR our side setting a contract) and the 36.5% is for any negative score (our side going down OR our side defending a making contract). While obviously any one negative score might be great (a good sacrifice) and any one positive score might be bad (missing an obvious game), on balance you can estimate very quickly and reasonably accurately your score just by looking how often the raw score is positive versus how often the raw score is negative.

 

That is totally separate from the our side declaring percentage (52.1%) or our side defending percentage (52.6%).

 

I'd be interested to know if anyone else tracks all their F2F match point results and if they see similar patterns.

[hotShot]

Playing matchpoints declarer and defender should get 50% if the normal contract is made.

At club level there are usually more that miss game/slam than those who overbid, so a "normal" declarer will get 50+X % for making, the defense will get 50-X % as a result.

 

So your expected value for the average declarer play should be 50+X,

while your expected average as defense is 50-X.

 

If you club has less than 7 tables and lots of weak players, than X might be quite big. At regionals and nationals that X should be smaller.

 

You get a positive score, whenever you make (even if you miss game/slam) or if they go down.

 

You get a negative score whenever they make or if you go down.

Defending, there is nothing you can do about the X when opps make.

You get a negative score and the best you can get is 50-X%.

You can get below 50-X by allowing unnecessary overtricks.

A good sacrifice will get you a better score than 50-X, a bad sacrifice will get you below 50-X.

 

What you need is an estimate of the X in your club. If that X in your club is 15, than the average score for a making game is 65% and the defense has to settle with 35%. Your result of 36.5% would indicate that you make a little more good sac's than bad ones.

But an X of 15 would indicate a very low level of play (or a very small club). If X is smaller, than you are making a little more bad sac's than good ones.

[jukmoi]

It is my experience that declarer gets on average about one trick over optimal result. I dont claim declaring is easy but defending is even harder. Shouldnt you start to bid more aggressively if your results declaring are positive and defending are negative?

[Mbodell]

What you need is an estimate of the X in your club. If that X in your club is 15, than the average score for a making game is 65% and the defense has to settle with 35%. Your result of 36.5% would indicate that you make a little more good sac's than bad ones.

But an X of 15 would indicate a very low level of play (or a very small club). If X is smaller, than you are making a little more bad sac's than good ones.

I also track, FWIW, the field strength of the games I play in (using the power ratings site). The average level of play of my field is around 4.5. And "average" game across the ACBL has a strength of 0. A bad field has a negative strength. Day 1 of blue ribbons is about a 9 and day 3 is about a 13. So the field in question is actually stronger than "average", but weaker than even day 1 of an NABC+ event.