Magic Contest:  Overview
 
 
Scoring:  Weighted scores (Law 12C1c)

 Home
 Magic Contest
    Overview
       Movement
       Enter results
       Enter contracts
       Double dummy
       Scoring
          Neuberg's formula
          Weighted scores
          Rounding
       Report contest
    Samples
    Documents
    Country versions
    Price list
    Download

Tomas Brenning
tomas@brenning.se
Tel:     +46 171 47 50 37
Fax:     ---
Mobile: +46 70 742 77 42

Veckholms-Åkerby 2
SE-745 99  Enköping
Sweden

Alternate e-mail address:
tomas.brenning@gmail.com

Law 12C1c will be implemented from version 2.10. I will let an extract from a document by Ton Kooijman describe what scoring with Law 12C1c means:


Beginning of quote

A weighted score exists when more than one result on a board is awarded to a pair, the scores being related with at probability of occurence. When the TD deems the chance to defeat 3N as too small to award a full 3N minus one but still considers it a real possibility to defeat it he could decide to award 2/5 of -100 and 3/5 of 400. And it could become more complicated. Due to an infraction a pair doesn't reach a game, the denomination not being obvious. And in 3N, with a chance of 30% to be reached, there will always be 9 tricks, and in 4S taking care for another 30% there is only a chance of 1/3 to make it. In the remaining 40% the TD supposes the pair to double the contract reached by the opponents.

To convert these scores to a result in matchpoints we add these scores to the frequency table one by one and calculate the matchpoints for each of these results. Then we multiply with the expectancy as estimated. The questions what to do with the scores of all the other participants is hardly touched. It is worse, I suspect that quite often those matchpoints are calculated with a score less, the weighted score. It would be an improvement to use the Neuberg formula but there is a much better solution.

Let me show what should be done. To keep it easy we assume a frequency table with only two scores: 2 times +600 for 3N just made and 5 times -100 for 3N minus 1. There are 8 tables and for the missing one the TD assigns a weighted score based on 40% making 3N and 60% going one off.

What we have to do now is to add these frequencies: 0.4 times 620 and 0.6 times -100 to the frequency table, giving the layout of the first score table to your right.

The matchpoints are calculated by adding 5.6 to -1 giving 4.6 and adding 2.4 and 5.6 to 4.6 giving 12.6. This is the basic method for calculating the matchpoints of any frequency table.

The matchpoints for the weighted score are calculated by taking 0.4 times 12.6 (5.04) and 0.6 times 4.6 (2.76) giving 7.8 (see the second score table to your right). This is of course the same result as adding both scores one by one to the frequency table for the seven scores: 3 scores of 600 give 0.4 times 12 = 4.8 and 6 scores of -100 give 0.6 times 5 = 3, which adds up to the same 7.8.

I am interested to know which software programme will be the first to calculate results this way. And for future EBL (WBF) events I consider it mandatory to do the calculation like this.

End of quote


I hope Magic Contest is the first!

Entry method

In Magic Contest's result window you enter weighted scores just as shown below, i.e. "12C1c:2/600/3/-100". This way you do not have to bother with percentages - just enter number of times the different types of contracts make.

Example


12C1c:2/600/3/-100

Basic score table with
frequencies from 12C1c included:

# Result   N-S E-W
2.4 600       12.6 1.4
5.6 -100   4.6 9.4


12C1c:2/600/3/-100

# Result   N-S E-W
2 600       12.6 1.4
5 -100   4.6 9.4
1 12C1c   7.8 6.2


12C1c:2/600/3/-100/5/-300

# Result   N-S E-W
2 600       12.8 1.2
5 -100   5.3 8.7
1 12C1c   3.9 10.1


(1) 12C1c:2/600/3/-100
(2) 12C1c:2/600/3/-100/5/-300

# Result   N-S E-W
2 600       12.4 1.6
4 -100   4.9 9,1
1 (1) 12C1c   7.9 6.1
1 (2) 12C1c   3.7 10.3