Sometimes you read something that makes you think “why would they do that?”. This time that happened to me when reading one of the bulletins from the WMG in Rijsel, #10 to be precise. To get the document, click here, the relevant section is on page 10, it reads:
- Change in the victory points scales
- A Scoring Committee was set up by the WBF President in January of this year. [...] After deep and long analysis, the committee proposed — and the WBF Executive Committee accepted — the following:
- 1The adoption of a 20-point victory point scale with the following features:
- The scales are continuous and given to two decimal places
- [...]
This looks quite innocent until you think about it for a second. Right now, team matches scored in VP use a step scale: 15-15 VP for a tie (that is a 0 to 3 imp’s difference over 24 boards), 16-14 VP for 4 to 9 imp’s difference, up to 25-0. This will be changed: 0 imp’s will be scored as 10.00 against 10.00, 1 imp as 10.25 against 9.75, 2 as 10.50 against 9.50, 5 imps presumably as 11.16 against 8.84 or something like that.
My first question was “what problem are they trying to solve?” The article isn’t clear about it, but it appears that the main reason for this proposed change is the perceived unfairness in the current system. Due to the discrete steps in the scale, not every extra imp translates into a bigger victory point score. If you win a (24 board) match by 57 imps, you have done better than a team in another match that wins by 56, yet, you both score 25 VP. Change your result to a win by 52 and 51 for the other team, and suddenly you score 25 VP and the other team only 24. With a continuous scale, in both cases, you will score a fraction more than the other team and thus be rewarded better.
This is true, though the counter argument is bridge-wise, there isn’t a real difference between winning by 57 or 56 imp’s. The imp scale itself is discreet, so there will always be small random effects in the score. A simple example: team A bids and makes 6♠ for 1430, their opponents go down, team A wins 17 imps. In another match, on the same board, team B bids and makes 6♠ with an overtrick for 1460 while again their opponents go down, for the same 17 imp win. Yet, team B has taken more tricks and thus arguably done better. Should they get 17.1 imp? If you follow the argument above, yes, but I doubt that anybody is considering to change the imp scale.
Another way to look at this, is similar to something that happened to me back in high-school. Pocket calculators were quite new, and my teacher proudly announced that I had scored 6.45 on some test. The kid next to me had scored 6.46 on the same test. I asked what I had done worse than him. The teacher couldn’t tell me, other than that this is what his calculator said. Of course, the real answer was that there was no difference, it was simply a rounding effect when adding up numbers with a small margin of error. The same is true for a 57 vs a 56 imp win, you have beaten the opponents badly but on the data available, it is impossible to tell who did the worst.
The next question that needs to be answered, isthe overall effect of this change. Sure, moving to a VP scale in 2 digits will change the individual match results. But what matters is the overall order of the teams, when all VP scores have been added up. In fact, this issues has been raised before and the previous committee who studied it, wrote: “there was practically no change in the ranking of the teams at the end of a round robin” (Ernesto d’Orsi, chair WBF rules committee in 2002).
And then there is the argument of publicity. The current 25-5 VP scale has been designed with publicity in mind. This one has not. To understand that, we have to go back in history. Originally, bridge used the same 2, 1, 0 scale as used in many other sports, 2 points for a win, 1 for a tie, 0 for a loss. Simple and clear. A win was defined as more than 1/8 of an imp won per board played, or to win a 32 board match, you needed to score at least 5 imp. Over time, this wasn’t considered fair, as one felt that a win by 100 imps was more significant than a win by 5. So, a 4-0 scale, a 6-0 scale and others were introduced until in 1968 a 20-0 scale was introduced.
At this time, also minus points were introduced, that after 20-0, it went 20-(-1), 20-(-2), until 20-(-5). (Where the 20-(-1) wasn’t universally used). The idea was that at some point, the losing team should be punished for playing badly, without further rewarding the winning team. That does make some sense. The 20-(-5) scale had one disadvantage though: the minus points were hard to understand for non bridge players and often got incorrectly published. In 1983, a simple solution was found: add 5 top all VPs and you got rid of them. Simple, problem solved. Suddenly you could tell your non bridge playing friends that you won a match 18-12. As the concept of deal exists in most card games, they’ll immediately understand that.
With this proposal, we’ll go back to something like 18.33 against 11.67. “What”, they’ll say? Right, is the perceived increase in accuracy really worth the decrease in understandability? Again, the previous time the scoring committee studied this, they concluded “the benefit of the greater accuracy in the results would not compensate for the complicated way in which the result would be presented”. (d’Orsi again)
Bottom line of all this, a simple discrete scale will be replaced by something that looks incredibly accurate but, in fact, suffers from the same rounding issues as the discrete scale. Life becomes incredibly more complicated without solving any real issue. I wonder why anyone would do that.
Finally, time for a commercial. On October 20, Gijs and Anita will organize a pairs tournament in their bridge home in Zwolle. For details, click on the logo.