Join Bridge Winners
Leveling the playing field (long)

What does the average ACBL member pay for a single session at a regional or NABC? When you include travel, hotel, meals and table fees, I'd guess the average would be north of $50, even for the most frugal of ACBL members. In light of this, the ACBL should do everything it can to ensure that the competition is fair.

Let's consider a typical case -- a 14 table Mitchell in which the boards are duplicated by the players. The movement is imperfect in that all pairs miss one set of boards and one pair of opponents. If you miss the strongest pair in the field you start out with a significant advantage over the pair that misses the weakest. Let's say that you duplicate boards 1 and 2 and that you are sitting E-W. You will miss two boards where you are NV, and therefore two boards where you have the greatest scope for action. Finally, if the boards themselves are uniquely suited to your methods, missing those boards is a disadvantage. Your chances to win the event could be affected by things over which you have no control.

It gets worse. To be perfectly fair, the N-S and E-W fields should be balanced in strength. In an ACBL tournament event, this is commonly done by distributing the strong pairs as evenly as possible, and letting the chips fall at random with respect to the weaker pairs. But distributing the weak pairs is just as important. Your fortunes are partly dependent on how many gifts you get, and weak pairs are the primary source of gifts.

In making seeding decisions the directing staff is faced with an insoluble problem: They don't know for sure what the final composition of the field will be like. If there will be four sections, then they need to determine who the eight best pairs are and distribute these eight accordingly. But they don't always know how many sections to expect, nor which strong pairs will show up. A very good tournament staff can perhaps do a creditable job with the top three pairs in every section -- or maybe not. At best, manual seeding is a hit-and-miss proposition because it takes place with incomplete information.

We decided long ago that purely random deals were appropriate for all serious bridge events. That's the nature of the game. There is, let us face it, a fair amount of luck involved. Doesn't it follow that we should strive to level the playing field as much as possible?

In a perfect world we would have an accurate method of ranking all pairs according to strength. I have experimented at the club level with the notion of ranking players by their average order of finish. This turns out to be a very good measure of overall strength, IMHO. It simply counts the number of pairs that the player has ranked above in an event with the number of pairs that the player has ranked below. With each event played, the number of "aboves" and "belows" are accumulated and player strength is simply indicated by the ratio of these two totals. Gathering the data could be done with a simple modification of the software.

This ratio is dependent on context, of course. The "big fish in a little pond" effect would dictate that data be collected in context. That is, a player would have one ratio in the local club, another in sectional tournaments, a third in regional tournaments, and a fourth in world class competition. Eventually, with enough data collection, these four measures could be correlated so that a player's strength in his or her local club could be used to predict their probability of success in higher ranked contexts. In theory, the ACBL would be able to rank any player against the Meckwells of the the bridge universe, even though they never played in the same event.

I am not by any means suggesting that we drop the masterpoint plan, which has proved so successful. I am suggesting only that these four ratios be used for the sole purpose of balancing fields in ACBL events. These ratios would not be a state secret, but would not be actively publicized either. The paradigm of player lifetime achievement as represented by masterpoints would not be abandoned.

I can tell you from personal experience that many players are acutely aware of the value of avoiding strong pairs and of playing weak pairs. If the vagaries of the seeding system dictate that they play all the strong pairs, while not playing many of the hopelessly weak pairs, a pair's chances of finishing high in an event are considerably diminished. This is glaringly obvious at the club level, where everybody knows everybody else. At higher levels, it is not so obvious; we don't always know who is strong and who is weak.

We must first acknowledge that bridge competition has a strong random element. If fate ordains that you must defend a slam contract that can be made with a simple squeeze, you are screwed if you play that board against Meckwell, and blessed if you play it against Mick and Milly Cardpusher. Organizers cannot remedy this inherent "unfairness", but they can do something to level the playing field. They can -- and should -- ensure that your opponents are a reasonable mix of strong and weak players.

Unfortunately, this is not possible when the tournament staff must make seeding decisions on the fly. Even if, as suggested above, tournament staff had an accurate measure of each pair's competence, in the few moments they have to make seeding decisions they could not possibly decide accurately how to seat every pair in the field optimally.

As an example, suppose we knew the field consisted of 56 tables -- 4 sections of 14 tables, and could accurately rate them by strength from 1 to 56. No human being, however talented and knowledgeable, could arrange these 8 fields (4 N-S fields and 4 E-W fields) such that the each field were as nearly equal in skill as mathematically possible, certainly not in the time available at an ACBL tournament. No human being is that good. Yet a silicon chip the size of your thumbnail, properly programmed, could do exactly that in microseconds.

Why would we not avail ourselves of such perfection? The programming effort necessary to bring it about would be trivial compared to the advances in fairness that it would bring. The answer is that with our current methods for registering pairs for an event, we cannot do this. How could we get every contestant's entry fee and ACBL number, enter those ACBL numbers into the program, and then inform every pair of their starting table and direction?

I think there is a way. Electronic bridge. (EB)

My idea is an iPad for each player (locked in a framework) with software that connects wirelessly to a central computer. Bridge actions are made via touch screen or mouse, at the player's choice.

OK, it seems radical, but consider the advantages:

Better Bridge Illegal actions would be impossible (No revokes, no bids or plays out of turn, etc.) Breaks in tempo would be all but eliminated as a thorny problem simply by not showing the last three calls until RHO has called. Alerts would be electronic and visible only to the opponents, eliminating another source of UI.

Security With the addition of screens, cheating would be virtually impossible. Every competitive deal played in the past year -- complete with timing -- could fit on one hard drive and be available for review by anybody. And of course, Big Deal would generate each deal on the spot, eliminating any possibility of prior knowledge.

More bridge With EB, timing would be perhaps 5:45 or at most 6:00 per board. (Jimmy Cayne does much better than this every day on BBO.) I would guess that 32 boards per session could be the norm, and could easily be played in about 3:30, even with time for breaks. I like 32 boards as the default, especially since it equalizes vulnerability.

Better management of slow play issues. The time taken by each player would be known to the second. Slow play enforcement could be automatic and totally fair. I believethat slow play penalties would become all but non-existent. By starting play on the next board promptly, extended discussion of the previous board (the chief culprit) would be curtailed. EB, by its very nature, would discourage slow play.

Fairer Bridge Pairs movements would be much fairer. Every pair would always play every board and every opponent, the holy grail when it comes to fair movements. The only compromise is that you might play one extra board or one fewer board against some opponents. As an example, to play 32 boards in a 15 table section, you would play 13 rounds of two boards and 2 rounds of three boards. The key idea here is that in every round the same boards (in numerical order) are played at every table. As long as the field is balanced, this levels the playing field as much as possible.

Because of the nature of random deals, some sets of boards will favor one pair or the other, and this element cannot be eliminated, nor should we try to. But by balancing the skill of all the the N-S players with the skill of their E-W counterparts, we can achieve a level playing field.

The best balance can be obtained only by waiting until the last pair has registered. Balance can be made even better by using the strength index described above, but using total masterpoints would also be quite adequate. Pairs would buy their entry, take a seat at any vacant spot, and enter their ACBL numbers. As soon as the last number is entered, the program would choose the movement(s), balance the field, and send every pair to their starting assignment. The entire process for event startup would take place in just a few minutes, without any human intervention.

Swiss team events would benefit immensely. Instead of taking 4 hours for a typical 4x7 event, a 4x8 event would easily finish in about 3:30. Since they are all playing the same boards, every team has the same scoring potential. Much fairer, and much easier on tournament staff. KO events would also benefit, but not nearly as much as Swiss events.

Important events like the Spingold, the Cavendish, and the Bermuda Bowl could be webcast in real time. The ACBL could perhaps find a revenue opportunity here. What would you pay to watch the Bermuda Bowl finals in a live webcast?

Cost is a major consideration. At today's prices, the capital cost per table would run about $2000, including an external battery pack and mouse for each iPad. Some talented hardware guy could probably reduce that somewhat. I calculate that all Regionals and NABC's could be covered with 1500 units, meaning an investment of $3 million or less. Of course, one could expect prices to go down as time passes.

A pilot program with just 40 units could test the waters by conducting five or so EB sessions in selected sectional tournaments, maybe three matchpoint events and a two session Swiss every week.

I notice that software people are heavily represented here. Can somebody give me a ballpark estimate of software development cost? I'd want the software to balance the field, matchpoint or IMP each board across the field, do the scoring and pairings for Swiss team events manage the slow play penalties and add the event to the online archive. Maybe a few more bells and whistles.

What do you folks think?

13 Comments
Getting Comments... loading...
.

Bottom Home Top