The Losers Trick Count

The following simple study demonstrated that the Losing Trick Count is not as good as using Shortage Points as published in my book "Acol from Scratch" published on Amazon (2014, 2015 and soon to be published 2016).

The Losing Trick Count of LTC is a method of hand evaluation that was first documented in 1910.   You can read all about it on wikipedia here.  It's still very popular today.  An alternative method is to use Goren Shortage Points where points are added to your high card points (Milton Work Count) to adjust for the ability to ruff.  Holding four+ trumps add 5 for a void, 3 for a singleton, 1 for a doubleton.  Hold three trumps add 3 for a void, 2 for a singleton and 1 for a doubleton.  Notice there are two ways to do this, just add shortages to the short trump hand or add for shortages in both hands.  I call the version that adds for shortages only in the short trump hand "Dummy Points" the version that adds for all shortages I call simply "Shortage Points".

You can read a "paper" published on Amazon called "Bid Using ESP" that studies both point counts and offers some minor improvements.

I have listed over 100,000 deals all of which were played in major competitions.  The "Losing Trick Count Dump.ods" above is an openoffice spreadsheet, you can download openoffice for free in a few minutes.  I can't load it on here using a Excel sheet because it's too big.  However I have also loaded it into a zipped csv file "Losing Trick Count Dump .zip" without the correlation formula so those of you with Excel and perform the same test as me by adding your own CORREL formulae.  This data was source from the vugraph project and included all major tournaments between 1955 and 2013.  Looking at hands that hold a suit fit I have calculated the LTC winners, Milton Work HCP, Shortage Points and Dummy Points.  I have then added formulae to show the correlation of these point counts with the number of tricks made.  In other words how good are they at predicting the number of tricks.  A correlation of 1 means they are always exactly right and a correlation of 0 means the point count is useless.  I have performed other more complicated studies that take account of the fact that the method needs to be right when game is likely and they lead to a similar conclusion.

I have produced this dump so that those who are not familiar with SQL code can look at the data manually and check its validity, if you find any errors please let me know!

The Results where as follows:

LTC Correlation  0.49

Shortage Point Correlation 0.55

Dummy Point Correlation 0.49 

HCP Correlation 0.46

Sample 101482 Deals

As you can see Shortage Points win hands down.  This is a highly significant result with such a large sample size giving little room for sample error.  Dummy points are no better than LTC but HCP is a little less correlated in suit contacts.

Losing Trick Dump Small.ods Losing Trick Dump Small.ods
Size : 4443.364 Kb
Type : ods
Losing Trick Data Losing Trick Data
Size : 1857.655 Kb
Type : zip