Back to Thomas's Bridge Fantasia

Hand Evaluation Articles

This is a collection of articles which explore hand evaluation using what was, at the start, a new technique. Due to the revolution in the ability for computers to solve full double-dummy problems, we can generate huge data sets and start to investigate the value of hand shapes and suit holdings in some objective fashion.

I also discuss the reasons to distrust double dummy data, and explain oddities in the data that may surprise you.

Articles

Single Hand Evaluator: My original article, which estimates how many tricks you expect to be able to take (double dummy), looking at your own hand.
Binky Points: An Additive Evaluator: This uses the data from the previous article to determine an evaluator which is "additive," that is, which is defined so that the sum of the value of your hand and the value of partner's hand approximates the number of tricks available.
After Partner Opens 1♠ and you have a fit: A rough draft with some raw data. When you have a fit after partner opens 1♠, how does that affect the Binky Points suit evaluation of your hand?
Misunderstanding Binky Points: A poster on the BBO discussion board think there are obvious flaws in Binky Points, but he is just misunderstanding the values.
Testing Binky Points: Comparing Binky Points to other evaluators.
2NT Openings: A ridiculously detailed look at 3-5pt hands opposite a strong 2NT opener, and when it pays off to be in 3NT.
The Stiff Queen: And Other Binky Points Oddities: Binky Points gives the value of a stiff queen as worth 0.25 tricks more than a small stiff in suit contracts, even though the single-hand evaluator gives a difference of only 0.08. What gives?
Aces and Spaces: Checking Binky Points: Examining balanced hands with four aces and no other high cards or spots.
Card Values for Three Notrump: Another approach to valuations - targeted for three notrump. Is this the revenge of Milton Work?
Experts and 4333 Patterns: In a discussion on Usenet, there was some usage of my data to contradict expert practice to treat 4333 patterns as negative for notrump play. This is an example of a bad use of my data.
Conclusions and Warnings: Some conclusions, and some attempts to avoid jumping to conclusions.
High Card Points versus Binky Points: Noting that one high card point is worth approximately 1/2 a trick in notrump and 1/3 a trick in suit contracts, and looking at 'refined' values for holdings from a HCP viewpoint.
An Interactive Evaluator: Allows you to try out various evaluators online.
Problems with using double dummy data: There are problems using double dummy data when building evaluators.
Problems with using single dummy data: Some people suggest using single dummy data as the answer to the problem of double dummy analysis. Single dummy data, however, has its own set of problems.
Cowan's 5-4-3-2-1 Result: Richard Cowan determined that trick values for honors down to the ten correspond fairly closely to 5-4-3-2-1. What's wrong with this picture?

Data Files

Raw Data

NT tricks taken, by holding
NT tricks taken, by hand pattern
NT tricks taken, by suit length
Suit tricks taken, by holding
Suit tricks taken, by hand pattern
Suit tricks taken, by suit length

Binky Points Data

New! I recently regenerated the Binky Points values from a bigger set of data. There were only minor difference except for very long suits where I had little data before.

Binky Value of Holdings [old data]
Binky Value of Patterns [old data]

Comments

I am going to try an experiment using Disqus to facilitate commenting here. Might work, might not. Keep it clean and fun, and I'll see if spam is an issue.