Release 2.19r

Books are available on loan to club members. Visit Library.

Let's Play Bridge
with Marx Brothers
Visit  a Bridge Village
in India.

"We all make mistakes when playing!"

Frequently Asked Questions
  Cross IMP Pairs

What is it all about and why do we do it?

1. How is Cross IMP Pairs scored?
2. What are the advantages?
3. What are the disadvantages?
4. Never mind that, why are we doing it?
5. How should we adjust our game with this sort of scoring?

1. How is Cross IMP Pairs scored?

Your cross IMPs on any particular board are calculated by comparing your score with everyone else who played the board in the same direction as you.  The total of these comparisons in IMPs is averaged to give you your cross IMPs on the board.
The cross IMPs for all boards are then added up. The total is adjusted if some pairs played fewer boards then others to bring it in line. 


2. What are the advantages?

a. In Cross IMP scoring you just need to make your contract to get a suitable reward, especially games and slams, making overtricks doesn't gain you much.

b. There is little to choose between 3NT and 4 of a major and the few points difference is almost irrelevant, so you should choose the safer game, the same as if you were playing rubber bridge.

c. At normal pairs scoring (matchpoints), pairs have a natural fear of going past 3NT in search of a minor suit slam. This has the effect of making people's slam bidding rather poor. As 5 of a minor will score OK compared to 3NT at IMP scoring, your slam bidding should improve.

d. Safety plays that risk not getting an overtrick in return for greater safety are a luxury that you (usually) cannot afford at matchpoints. Yet a lot of bridge literature is about this very topic and may well have figured significantly in any lessons on the game you may have taken. It is nice, at least for a change, to have a form of scoring where safety plays may figure more prominently.


3. What are the disadvantages?

a. Matchpoint bridge is still exciting, with tactics all of its own.

b. Matchpoints are simple to score and simple for everyone to see where their score has come from.

c. At matchpoints, as each board is worth a simple top, bottom or a shade of grey in between, over a limited number of boards (and 24 is limited), you may get a fairer result (fairer in the sense of less random) as each board is worth the same amount. If a good pair comes to your table and bids a slam that the rest of the room has not been able to find, you will get a bottom at matchpoints which may be no fault of your own. This is an unfortunate setback, but because each board is worth the same as every other one, you can potentially recover on the next (perhaps partscore) board. At cross IMP scoring you will need at least two, probably three partscore swings in your favour to make up the lost ground. This is not a concern in the long run (luck will even itself out), but in a single session at cross IMP scoring a few 'big boards' has a bigger effect than at matchpoints.

This is probably the main reason why matchpoints are the norm for pairs scoring for most sessions - though it is not a compelling reason why all sessions should be scored that way.


4. Never mind that, why are we doing it?

Well, apart from the things listed above as advantages of cross IMP scoring, the actual history goes back to the last Thursday of the month when we had 'multiple teams'. Some did not like teams because they did not like potentially letting their teammates down. Some didn't like them because sometimes moving around the room was complicated and also they didn't know how they were doing as scoring units don't show percentages. Cross IMP pairs seem like an ideal compromise. There are no teammates to let down and moving around the room is like for any pairs event. Our scoring units will show cross IMPs (not percentages) and this gives you an idea how you are doing - 0 cross IMPs is like 50%, any + score is more than 50% and any minus score is less than 50% - roughly.

Additionally, there are those who play team matches outside the club sometimes, but, as nearly all bridge is pairs (and therefore matchpoints), there is actually very little opportunity to practice the tactics required for team matches. Playing at least some sessions of cross IMP pairs provides that opportunity.

Also it is a bit of variety, it is nice to have a different form of competition sometimes in order to try to give something to everyone. Cross IMP pairs is just part of that overall effort.


5. How should we adjust our game with this sort of scoring?

You could write a book about this, but briefly:

a. When choosing which game to be in, you should choose the safer option.

b. You should not fear being in 5 of a minor so much as you do at matchpoints, especially if you want to investigate the possibility of slam.

c. If you think your side has the majority of the points, but not a game, then if the opponents bid over you, you're in a situation where you will need to think if you should bid one more yourself, double the opponents or just pass. This is a very common situation. Often at matchpoints you should double and pass is commonly the wrong answer. The reason is simple enough - if the opponents can make their contract you're getting a poor score anyway, so the extras they get for making a doubled contract won't hurt you very much, if at all. Whereas an extra 50 or 100 if you can get them off may make quite a difference to your matchpoint score. At cross IMPs the maths is a little different - the extras they get for making a doubled contract, particularly if you double them into game, will hurt you badly. And, if they're off, the extra 50 or 100 won't benefit you very much. In any case, at cross IMPs, if it is a partscore board that your side should be getting a plus for, it will, at least, not be a bad board for you if you collect any sort of plus at all.

d. At matchpoints, if you have a close decision between bidding game or not, then it is generally right to be conservative in your approach. The reason is simple, if the game turns out to be worse than a 50/50 bet then the advantage of staying low is obvious. And if it is a 50/50 bet then, by staying in the partscore, you will still pick up a few matchpoints from those who have got into the wrong denomination (which is more common than one might think), provided, of course, that you don't get too high yourself. At cross IMP scoring the game bonus is worth pursuing especially vulnerable and you should be more aggressive in your game bidding (marginally).

e. At cross IMP scoring generally you should play to ensure your contract as overtricks are of little value. And in defence you should play to get the contract off even if this risks giving away an overtrick.

Modified for Chislehurst from the original article here.

  Computer Dealt Hands

Since we started using the card dealing machine, there are a lot of weird distributions, long suits, singletons and voids.
Are the hands truly random or are they fixed? What has happened to our normal hands?

These funny computer-dealt hands!!
Well, hand-dealt boards tend to, if not shuffled properly, give more flat distributions than statistically probable. That's because cards of the same suit end up next to each other when the cards are played or when tricks are gathered together. If these aren't separated by shuffling, it ends up that one card in that run of cards of the same suit goes to each player during the deal: voids and singletons become less likely.

When a card dealing machine is used, voids and singletons occur with the correct probability. Hands will now look more distributional, especially if manually dealt hands didn't use to be shuffled thoroughly.

Also see Bridge Probabilities from Wikipedia.
This says, for example, that the probability of getting a 7 card suit is 7 in 200 hands, ie for every 200 hands you get 7 card suit 7 times - on average. Six card suit turns up much more often - about 17 times in 100 hands. Also it seems that singletons are quite common - 3 times in every 10 hands, while on average there are 5 voids in every 100 hands. And - if you think that flat hands are now very rare, perhaps you would change your mind if I tell you that - on average - there are 47 flat hands in every 100 hands. Sometimes more, sometimes less.
Read also the following:

"Since we started using computer dealt hands, I've been extracting statistics from the hand definition files - the PBN files - and fed them into an Excel spreadsheet. Data from 2007, 2008 and 2009 have been used, 18048 deals in all. The results follow quite closely the theoretically expected values - see Summary.
I also produced a series of graphs which demonstrate random behaviour from set to set. The process has been semi-automated, so it didn't require as much work as one might imagine and it can easily be updated with more data or repeated for another set of data. The spreadsheet itself is also available for download - beware, it's size is about 7Mb.

High Card Points Graphs
Singletons Graphs
Suit Breaks for 7 and 8 card fits
Suit Break Graphs for 9, 10 and 11 card fits
Hand Patterns Graphs - 1
Hand Patterns Graphs - 2
Hand Patterns Graphs - 3
The Excel Spreadsheet

1-March-2009 " 


  What is Par Score or Optimum


  • On Bridgewebs hand diagrams in the top right hand corner it says Optimum  followed by contracts and/or scores.
  • On our own style hand diagrams in the top left hand corner it says Par Score followed by contracts and scores.

What does this mean?

That is the score where neither side can improve their score by bidding further.


  Makeable Contracts and Points on Hand Diagrams

On the results page each hand diagram shows a framed square in the bottom right hand corner, with some numbers in it. What are the numbers?
Also, what are the four numbers in the bottom left hand corner?

The numbers in the bottom left hand corner are the high card points for North, South, East and West.
The numbers in the bottom right hand corner show makeable contracts. So, if there is number 6 to the right of N(orth) and underneath the 
 symbol, this means that small slam in hearts is makeable on this hand, played by North.


  When Top is Not 100%

We got an outright top on a board and yet on our personal score card it is shown  as 99.54%. Why not 100%?

When all boards are not played the same number of times, the match points for the boards played fewer times are factored/increased to allow for this. The Neuberg formula is used to do the factoring. However, the effect of the formula is also to 'downgrade' the match points on these boards a bit to allow for the fact that the 'competition' wasn't as good as on the boards played more times.

The calculation is performed as follows:

On each traveller match points are allocated to each pair first. This is done by comparing this pair's score with all the other scores on the board. (Adjustments/averages and missing scores are excluded from this comparison.) 2 match points are allocated for each score that is worse than this pair's score and 1 match point for each score equal to it.

Therefore the maximum number of match points anybody can get on a board is:

Top = (2 * Number of Scores on the Board) - 2

Number of Scores on the Board excludes adjustments/missing scores.

So if there are 12 scores on a board and one of them was an adjustment/average, then the top score on this board will be 20, while with 12 'normal' scores, the top score would be 22.


The Neuberg formula is then applied to the resulting match points.

The Neuberg formula is:
FMpt = 
(Mpt * MaxS) + (MaxS - ActS)

FMpt = the factored match points to be given on this board to a specific pair
MPt = the original match points given to the pair as calculated above
MaxS = the maximum number of scores on any board in the event
ActS = the actual number of scores on this board. This excludes adjustments/averages and missing scores.

So, for example, if a pair gets 12 match points on a board that was played 11 times and the maximum number of times any board was played is 12, then this pair will get 13.18 match points after the Neuberg formula was applied or 59.9% (13.18/22) and not 60% (12/20). Equally if a pair gets a top on a board played 11 times and the maximum number of times any board was played is 12, then this pairs score will be 21.9 match points after the Neuberg formula was applied and not 22, which would be the top score for the board played 12 times. In percentage terms this would be 99.54% (21.9/22), and not 100% (20/20).


Wish you hadn't asked?