Grands at IMPs?

[nigel_k]

Assuming the other table plays in the small slam:

 

Vul minor slam: If you make grand you gain 13 IMPS and if not you lose 16. So you need 16/29 which is 55.2%

 

Vul major/NT slam: If you make grand you gain 13 IMPS and if not you lose 17. So you need 17/30 which is 56.7%

 

Non Vul: If you make grand you gain 11 IMPS and if not you lose 14. So you need 14/25 which is 56.0%

 

 

Assuming the other side plays in game, the numbers are:

Vul minor: Gain 4 vs lose 12 so 75%

Vul major/NT: Gain 4 vs lose 13 so 76.5%

NV minor: Gain 3 vs lose 10 so 76.9%

NV major/NT: Gain 3 vs lose 11 so 78.6%

 

Therefore the percentage you need is somewhere between about 56% and 77%, depending on your assessment of how likely it is that at least a small slam will be reached at the other table.

 

This is for a long match. In a short match you may be slightly more aggressive. You also need to consider whether it may be a five or seven hand when you can have much less. And if the other table might not even play in the correct denomination you need more.

[Fluffy]

A bit of topic, but there are other factors, including: the opponents are saving and we aren't getting rich doubling, it is a 5 or 7 hand, or grand will probably be 50%, but partner might have some extra's I couldn't ask about

 

My last 5 grands at IMPS:

 

5% aprox to make, made

50% to make (from partner's point of view it was 1 of 3 good things happening, after I failed to give him the help with my shape or the key card, all that was left was a finese), made

50% to make, doubling them would had been 1100 for -8 IMPs, made

40% to make, but it was a 5 or 7 hand and we were already in 6, made

100% to make, 14 top tricks on 7NT against any lie out, including 9 diamond tricks, surprisingly grand was missed at the other table.

[mikeh]

If you want to maximize your VP expectancy in short matches then I think you should bid grands a bit more aggressively. I also think that it is a waste of time thinking about it.

 

To Phil: why do you think it does not matter?

I never said "It doesn't matter", but perhaps someone can explain to me why it would make sense to ever put ourselves in an EV- situation by not bidding a % grand?

The longer the match, the less likely it is that any single board will make a difference. In a short match, one game or slam decision can easily be the difference between winning and losing. In an long match, if you have one disaster, you have lots of other opportunities to make up for it.

 

Conversely, if you think you're behind in a short match, you'll need a big pickup to make up for it. So you might want to go for a big swing on a later board. Of course, there are two ways to swing a board: bid aggressively to a higher scoring contract than you think the opponents will bid, and hope to make it; or avoid a contract that the opponents are likely to bid, and hope they go down while you make. But a part-score plus an undertrick doesn't score as much as a game or slam bonus, so desperate players usually go the aggressive route.

I know this is a popular belief, but I don't share it.

 

Few pairs will consistently play 7 perfect boards in a row...even those capable of doing so for extended periods will screw up on occasion.

 

And few of us regularly play against that level of competition anyway.

 

So if we screw up at our table, it is wrong to automatically think that we are stuck....my view is that the other table owes us enough to cover at least one disaster and sometimes two....while when we are hot or the opps are having problems, we're the ones covering our teammates.

 

To start swinging because one feels one is down 10 or even 20 imos with 2 boards to go runs a serious risk of snatching defeat from the jaws of victory, while also maybe seriously demoralizing your teammates, who had just played a stellar set.

 

I feel this is true even at win/loss swiss but it is even more true at VP swiss....where the only justification for swinging is if you need some blitzes to stand a chance of winning, and where the whole team feels that finishing 2nd is the same as finishing 15th.

[billw55]

Shouldn't you already know if you play key card? Asking for the trump queen (or extra length i.e. guaranteed 10 cd fit) after 5c/5d response, using next step that isn't a signoff in 5M. This is basic, standard after key card

Like I said, I know very little about bidding. So if 5♦ after 5♣ asks for trump Q, what are the answers?

[gwnn]

first step no

second step and above: yes and shows something else (a King or maybe a Queen if you know for sure where the king is)

[jjbrr]

i play retreat to our suit=no, 6 our suit=yes but no kings, 6 anything else=yes Q and also the K in this suit.

[kfay]

i play retreat to our suit=no, 6 our suit=yes but no kings, 6 anything else=yes Q and also the K in this suit.

+1

[gwnn]

what is 5c-5d, 5h and 6h?

 

also some people without vertigo start spiral scans, you bid the first king you dont have in some order.

[phil_20686]

If you want to maximize your VP expectancy in short matches then I think you should bid grands a bit more aggressively. I also think that it is a waste of time thinking about it.

 

To Phil: why do you think it does not matter?

I never said "It doesn't matter", but perhaps someone can explain to me why it would make sense to ever put ourselves in an EV- situation by not bidding a % grand?

well, there is one obvious theoretical reason:

 

If i am slightly ahead in a short match then the roughly logarithmic VP scale could mean that losing 13 loses significantly more VPs than winning 13. Similarly if you are behind you stand to gain more than you lose. In practice its hard to know whether you are ahead or not, but, practical difficulties never stopped anyone discussing it on the forum right ? :rolleyes:

 

 

Also, as an aside, I am constantly missing grand on those hands where grand is identical to small, sometimes they are both on the same finesse, or both on bringing in a 9 card fit missing the Q. I think this is important as the small slam is seldom completely cold and the less good it is the less good grand needs to be. There are a huge number of slam hands where the extra tempo of a bad trump break gives you an extra temp to cash a side suit winner and beats small when grand is easy on a reasonable layout.

 

I can even remember a hand where 3N, 6N, and 7N all required the same finesse to work - the only difference being how many you went off if it was wrong. (although 5m was legitimately cold i think).

[hanp]

perhaps someone can explain to me why it would make sense to ever put ourselves in an EV- situation by not bidding a % grand?

 

I'll try to explain why an IMP %-grand is not exactly the same as a VP %-grand. If that's not the answer you are looking for, my fault for not understanding your question.

 

Of course others are right that this depends entirely on the results of your other boards, and if some of these boards have not been played the issue gets quite complicated. If you expect to win the match or the event anyway, you might prefer not to make large random swings with the other table. Problem is that you don't know what the other table will do, so I'll try to avoid these psychological issues as much as possible.

 

Let us assume that after all but one hand, the score is fairly close (say 0-10 IMP difference, all scores about equally likely), that we are not vulnerable, and that the swiss 20-VP scale commonly in use in the US is being used http://forums.bridgebase.com/index.php?act...=40180&p=475428. The last hand will be a small or a grand slam, and the other table will be in it exactly when we are not. Which odds do we need to make bidding the grand attractive?

 

The IMP odds we know now, we need about 56% as we win 11 IMPs if the grand makes while we lose 14 IMPs when the grand goes down. (If the other table might not be in slam then the necessay odds get lower and if the small slam can also go down then the odds get higher).

 

But this game is about finding the best IMP odds, it's about making the best VP odds.

 

Just summing up the VP's you win from the table you see that on average, we win 4.75 VP if the grand makes, while we lose 5,67 VP's if the grand goes down. This means that we should bid the grand if the odds exceed 55.3%, a little bit less than the 56%.

 

Hopefully this is a convincing argument that on average one should bid grands slightly more aggressively in short VP matches.

 

I must admit that my bidding is not yet good enough to notice the difference between 55.3% and 56%, which is why I wrote that it is a waste of time worrying about it. And then you started asking impossible questions and I wasted my time.

[Phil]

HanP, this is the second time in this thread you have failed to read/understand or just ignored what I wrote.

 

Instead, you gave a nice analysis about IMPs / VPs and the grand % required to bid both. While this is mildly interesting from an academic standpoint, I would agree its close enough not to worry about. And if I was really interested in calculating the VPs and converting that into a grand %, I could have handled that one on my own.

 

What I was hoping for was an analysis in response to my original question:

 

why it would make sense to ever put ourselves in an EV- situation by not bidding a % grand?

 

A sample position I was thinking about presenting to the math folk was this:

 

- We are in a match with a weaker team. Lets say we are, on average, 1 IMP / board better.

- The weaker team will NEVER bid the grand, but will ALWAYS bid the small slam.

- We are vulnerable, so we will win 13 if we make the grand and lose 17 if we do down.

 

As others have indicated, the longer the match, the better our chances, but it seems as though there is a relationship between the % required to bid the grand slam and the length of the match, and the result can be plotted on a graph. The VP result might matter for the output, but for simplicity, lets assume its a straight win/loss. I suppose you would like a % of our chances of winning the match and lets say I want 95%, but I'd be interested in some of the conclusions here as well.

 

As a math layperson, I compare this to a billionaire coming in and playing one hand of blackjack for a VERY large sum of money. Say the house has a 52/48 edge over me, which is a lot lower than what is required for the grand. If the casino can keep him there for 'x' hands, they might take the action. Or, they might take the action on a game that has a higher return, like the Big Wheel or Keno.

 

Here's a study Jeff Goldsmith put on his website a number of years ago, which also relates to some of the questions asked:

 

Study of IMP matches

[JLOL]

Mike,

 

Should the above vary based upon the length of the match?  In a short match (i.e. six boards), I would expect that it would need a higher percentage of success in order to actually bid the grand.

 

Yes? No?

Why would you think this matters?

 

If you want to maximize your VP expectancy in short matches then I think you should bid grands a bit more aggressively. I also think that it is a waste of time thinking about it.

 

To Phil: why do you think it does not matter?

 

I never said "It doesn't matter", but perhaps someone can explain to me why it would make sense to ever put ourselves in an EV- situation by not bidding a % grand?

 

I'll try to explain why an IMP %-grand is not exactly the same as a VP %-grand. If that's not the answer you are looking for, my fault for not understanding your question. [snip]

 

HanP, this is the second time in this thread you have failed to read/understand or just ignored what I wrote.[snip]

 

LOL

Edited July 7, 2010 by JLOL

[hanp]

HanP, this is the second time in this thread you have failed to read/understand or just ignored what I wrote.

I only posted 3 times and the first time was before you posted. How much more can I do?

[Phil]

(snip)

LOL. How cute.

[billw55]

OK I think we covered the odds and match equity situations pretty well B)

 

I think my other question in the OP got lost. Are there any stats out there on how often grands go down in actual play at high levels? Failing that, can someone who plays at such levels offer anything anecdotal?

[NickRW]

What I was hoping for was an analysis in response to my original question:

 

why it would make sense to ever put ourselves in an EV- situation by not bidding a % grand?

 

A sample position I was thinking about presenting to the math folk was this:

 

- We are in a match with a weaker team. Lets say we are, on average, 1 IMP / board better.

- The weaker team will NEVER bid the grand, but will ALWAYS bid the small slam.

- We are vulnerable, so we will win 13 if we make the grand and lose 17 if we do down.

I'm not sure you want a mathematical answer really - I don't feel like trying to provide one anyway.

 

I think what you're saying is that

1) You're the better team

2) Therefore you're likely ahead already

3) Therefore why put yourself in a position where you're risking a 17 imp loss when you can just cruise in the small slam

 

There is, of course, merit to that argument. OTOH:

 

1) Why do think you're the better team - is it not because you're better at these tricky decisions - if you don't put your better judgement to use then perhaps you're not that much better! Perhaps you could have done without this swingy board at the end of the match - but tough - you're in that situation now.

2) Weaker players don't NEVER bid 7 - nor do they ALWAYS bid 6 - those are both unsafe assumptions IMO

[Phil]

What I was hoping for was an analysis in response to my original question:

 

why it would make sense to ever put ourselves in an EV- situation by not bidding a % grand?

 

A sample position I was thinking about presenting to the math folk was this:

 

- We are in a match with a weaker team. Lets say we are, on average, 1 IMP / board better.

- The weaker team will NEVER bid the grand, but will ALWAYS bid the small slam.

- We are vulnerable, so we will win 13 if we make the grand and lose 17 if we do down.

I'm not sure you want a mathematical answer really - I don't feel like trying to provide one anyway.

 

I think what you're saying is that

1) You're the better team

2) Therefore you're likely ahead already

3) Therefore why put yourself in a position where you're risking a 17 imp loss when you can just cruise in the small slam

 

There is, of course, merit to that argument. OTOH:

 

1) Why do think you're the better team - is it not because you're better at these tricky decisions - if you don't put your better judgement to use then perhaps you're not that much better! Perhaps you could have done without this swingy board at the end of the match - but tough - you're in that situation now.

2) Weaker players don't NEVER bid 7 - nor do they ALWAYS bid 6 - those are both unsafe assumptions IMO

Nick it seems to me that we need to hold a few things constant for a calculation, which is why I specified 'always' and 'never'. I didn't specify when the board was being played but if you want to make it the last board, then OK.

 

Even under your set of assumptions, it seems that one of the reasons we are the better team is our ability to bid excellent grands. If during most of the match we are bidding our 40% vul games in usual fashion, and things aren't going our way in that department, we very well may need the grand.

 

And yes, I am looking for the mathematical answer if anyone is up to it.

[zasanya]

What I was hoping for was an analysis in response to my original question:

 

why it would make sense to ever put ourselves in an EV- situation by not bidding a % grand?

 

A sample position I was thinking about presenting to the math folk was this:

 

- We are in a match with a weaker team. Lets say we are, on average, 1 IMP / board better.

- The weaker team will NEVER bid the grand, but will ALWAYS bid the small slam.

- We are vulnerable, so we will win 13 if we make the grand and lose 17 if we do down.

I'm not sure you want a mathematical answer really - I don't feel like trying to provide one anyway.

 

I think what you're saying is that

1) You're the better team

2) Therefore you're likely ahead already

3) Therefore why put yourself in a position where you're risking a 17 imp loss when you can just cruise in the small slam

 

2) Weaker players don't NEVER bid 7 - nor do they ALWAYS bid 6 - those are both unsafe assumptions IMO

Nick it seems to me that we need to hold a few things constant for a calculation, which is why I specified 'always' and 'never'. ..........

 

Even under your set of assumptions, it seems that one of the reasons we are the better team is our ability to bid excellent grands...................

 

And yes, I am looking for the mathematical answer if anyone is up to it.

Actually when I play a team match in which I believe my team is weaker I will bid a grand more aggressively rather than NEVER.

If I am up against a definitely weaker team who NEVER bid a grand and against whom i expect to win 1 imp /non grand slam deal then it seems mathematically correct never to bid grand against this team as i expect to win by at least 7 imps in the other 7 boards in a short 8 board match.This of course assumes that you do not mind the margin of winning.

If you want to extract maximum VPs then thats a different problem which cannot be solved without putting in other variables like how many vps you want.

[barmar]

So if we screw up at our table, it is wrong to automatically think that we are stuck....my view is that the other table owes us enough to cover at least one disaster and sometimes two....while when we are hot or the opps are having problems, we're the ones covering our teammates.

This is a good point. How many times have you gone back to your home table, apologizing to your teammates for losing the match, but they smiled and said "We got you covered", and you end up winning by 10 IMPs? You never know what's happening at the other table; unless you're playing against Meckwell-class opponents, they're just as capable of screwing up as you are.

 

So unless you really know your opponents, it's pretty difficult to estimate the "state of the match", and probably a bad idea to change your strategy based on it; you may just be digging yourself deeper when you didn't even need to.

 

But regardless of state of the match, the impact of a big swing on a short match is still quite significant.

[barryallen]

Mike,

 

Should the above vary based upon the length of the match?  In a short match (i.e. six boards), I would expect that it would need a higher percentage of success in order to actually bid the grand.

 

Yes? No?

Why would you think this matters?

The greater the sample, the less the variance in the calculated odds, so the number of boards is significant. If you were tossing a coin 1,000,000 times it would not be a surprise that after 100 throws you came down with 60/40 and 500,020/499,980 after the million, although your odds on every throw will still be 1/2. For very small samples it pays to be overly aggressive if the intention is to win, less so if you are considering the percentages or where there is money involved.

 

 

These bridge odds are based upon you making the decision based upon the same percentages time and time again. Start changing your selection criteria throughout and it changes the original odds.

[campboy]

Assuming the other side plays in game, the numbers are:

Vul minor: Gain 4 vs lose 12 so 75%

Vul major/NT: Gain 4 vs lose 13 so 76.5%

NV minor: Gain 3 vs lose 10 so 76.9%

NV major/NT: Gain 3 vs lose 11 so 78.6%

There's a factor of two error here. If you bid the grand and it fails you do score -11 on the board, but that is a loss of 22 compared to the +11 you'd get for making the small. So 88%.

[dburn]

I was hoping that some of the actual mathematicians would refer to the Parabolic Utility Function (oft quoted by Jeff Rubens, an actual mathematician, in the context of what to do in "sufficiently short" or "sufficiently long" matches).

 

Maybe they did, and I am too stupid to know that they did. But I still wish I knew what it meant.

[barryallen]

I was hoping that some of the actual mathematicians would refer to the Parabolic Utility Function (oft quoted by Jeff Rubens, an actual mathematician, in the context of what to do in "sufficiently short" or "sufficiently long" matches).

 

Maybe they did, and I am too stupid to know that they did. But I still wish I knew what it meant.

Just go and look at the results of tournaments on BBO where you have a set of 8 boards played by 50 pairs. The chances of you winning by adopting standard percentages are minimal, although your percentages will be well in.