ACBL Seeding Policies: a Vicious Cycle
(Page of 4)

The Spingold and Vanderbilt are seeded events. A computer determines Seeding Points from a formula and then, with some input from a Seeding Committee, sets the seeds and creates the brackets. However, the current system is badly flawed, with a bias toward older players and against younger ones. In this article, I will describe the current system, explain why I think it's flawed, and then propose a system that I feel is more equitable.

The current seeding system

Currently, only two factors go into a player’s Seeding Points: masterpoints and recent success in the three major ACBL team events — the Vanderbilt, the Spingold, and the Reisinger. The masterpoint component is based on a logarithmic formula, with a maximum of 11 Seeding Points based on masterpoints. (The maximum number of Seeding Points any player can have is 50.) You need a little over 20,000 masterpoints to reach 11 masterpoint Seeding Points. The rest of the seeding points come from results over the last ten years in the Vanderbilt, Spingold and Reisinger. The most recent results are weighted more, and the results-based Seeding Points “decay” each year: 10% the first year, then 11%, then 12.5%, on up to 50% by year ten.

Foreign players are assigned Virtual Seeding Points and Virtual Masterpoints by the Director in Charge (with assistance from the Seeding Committee). The Virtual Seeding Points decay at 10% a year for ten years; the Virtual Masterpoints are added to actual masterpoints and this total is used in the logarithmic formula.

An alternative method uses WBF Masterpoints and WBF Placing Points; whichever formula nets a higher total — the normal method or the alternative WBF method — is used.

The full details are available here: http://www.acbl.org/assets/documents/play/Conditions-of-Contest/Appendix-B.pdf

Problems with the current system

There are so many problems with this system that it’s difficult to know where to begin. But let’s start with my main objection, that this system discriminates against newer players. We all know thatmasterpointsare a poor barometer of bridge skill; usingmasterpointsmostly won in the seventies and eighties to predict a player’s chances in this year’sSpingoldis ludicrous. Consideringmasterpointsmakes sense when seeding the bottom half of the field, which is not likely to have many impressive showings in the Vanderbilt orSpingold; using them for the top half is a disservice to everyone but the old-timers.

Considering results in only three events makes the seeding process cyclical. The point of seeding an event is to give the higher seeds an easier draw in the first few rounds. While everyone will tell you that nowadays there’s no such thing as an easy draw in the round of 32, some draws are easier than others. When youoverseeda team to begin with, that error gets compounded when they get an easier draw and have an easier time advancing; now they have an easier road to winning more Seeding Points and cancelling out the “decay” of their previously earned points.

It’s difficult to win a match in the Vanderbilt orSpingold. It’s especially tough if you don’t have any Seeding Points, as you’re up against top teams in the first round. When you lose that match, you get no Seeding Points, and you’re in the same spot next year. The fact that the look-back period is ten years further works against newer players who do not have ten years’ worth of results. Someone who was playing ten years ago — even if not at a very high level — has an advantage because he at least has some results to add to his total. Even if he just won one match each event (which was much easier back then) he’s way ahead of the player entering the event for the first or second time. (Just for reference, advancing to the round of 32 in both theSpingoldand Vanderbilt in each of the last ten years is worth 5.5 Seeding Points.)

These problems have been improved somewhat by a rule added a few years ago that allows the Seeding Committee some leeway to use its judgment. Once the Seeding Points have been calculated, the top 16 seeds are fixed. Teams 17-32 are then placed into four-team groups — 17-20, 21-24, etc. — and teams 33-64 are placed in eight-team groups. Teams 65 and higher are placed into 16-team groups. The Seeding Committee can move a team up or down one group. (Exception: it cannot move a team from the 17-20 group up into the top 16.) When a team is moved into a new group, a team from that group must be moved up or down so that the groups retain the same size. Once the Seeding Committee has set the groups, the groups’ members are shuffled and randomly filled into the seeds assigned to that group.

This gives the Seeding Committee a chance to consider results other than just the Big Three team games; if you’ve won the Bermuda Bowl, they can move you up one group. But even if you’ve won the Bermuda Bowl, if you haven’t had much success in the Big Three events and start as the 64 seed, all the committee can do is move you up to the 49-56 group. That’s better than playing MONACO orNICKELLin the first round, but it’s probably not where you really should be seeded.

If the formula for determining Seeding Points is going to remain so flawed, the Seeding Committee needs more discretionary powers. It needs to be able to change the top 16. And it needs to be able to move people around more. Basically, it needs to be able to seed the event. It can use Seeding Points as a starting point, but it needs the power to use its discretion when the seeding formula is ineffective.

We can't lock ourselves into a seeding formula that is so flawed. Discretionary seeding wouldn’t be nearly as important if the initial formula were better.

Proposed changes to the seeding formula

The first change I would make would be to change the waymasterpointsare counted in determining Seeding Points. First, using lifetimemasterpointsis a poor method of predicting current success, becausemasterpointsearned earlier are not nearly as indicative of skill asmasterpointsearned recently. So I would use recentmasterpoints, say,masterpointswon in the last ten years. Probably Platinum Points won in the last ten years would be a better yardstick than just straightmasterpoints.

Second, I would include other events. The idea behind the current system is that the Vanderbilt,SpingoldandReisingerare the premier events, and that’s all we should be counting. There’s a lot more to bridge than these three events. TheACBLhas three major pair games each year — the Platinum Pairs, theLMPairs and the Blue Ribbon Pairs. These are counted asNABC++ events, just like the Big Three; they should be included in calculating Seeding Points. Sure, they’re a different form of scoring, but so is theReisinger. It’s still bridge. It’s difficult for an unestablished pair to do well in the Vanderbilt orSpingold; between finding good teammates and getting seeded high enough to get a decent draw, they’ve done well if they win one match. As random as pair events can be, success is much more in the pair’s control. Pair events are seeded, but a low seed doesn’t force you to playMeckwellall day. And you don’t have to worry about finding teammates; you just show up, pay your money, and play bridge.

There are other major events that should be counted. The US Team Trials has a weaker field than the Vanderbilt orSpingoldbecause the top foreign teams are not there, but it is a strong event, and success there should count for something. World Championships should count — not instead of other results, but in addition to them. The European Championships, other zonal championships. US players are actually at a disadvantage here: when a foreign player enters the Vanderbilt orSpingold, his entire bridge record is considered when he is assigned Virtual Seeding Points. When an American enters, it doesn’t matter what else he might have won, all that counts is the Big Three. The Seeding Committee can move him up a group, but that’s it.

There’s a problem with counting these other events: what happens to people who don’t play in them? We don’t want to just reward people who have the opportunity to play all the time. Most of us don’t qualify for World Championships, and we’re not all travelling the globe to play in every zonal championship. Even with theNABC++ pairs games, two of them are scheduled opposite otherNABC+ events. (The final rounds of theGNTare held opposite theLMPairs and the IMP Pairs is opposite the Platinum Pairs.) It doesn’t seem fair to punish players who play in these other events, or who can’t make it to the European Championships. Just like it doesn’t seem fair to punish the new pair that has never played in theSpingoldand so doesn’t have ten years’ worth of results. Easy fix: instead of just assigning a number of points for each result and adding them all together, you calculate the percentage of available points won. Let’s say we decide that winning anNABC++ pair event is worth 5 Seeding Points. At the end of the year I have 6 total Seeding Points from these pair games. If I played in all three, I won 6 out of a possible 15, so that’s worth 0.4. (These numbers don’t mean anything yet; I’m just giving a concrete example.) If I only played two events (my team was still in theGNTs, say), Iwon 6 out of a possible 10, for 0.6. There would need to be some minimum number of events counted; you don’t want someone to just enter theLMPairs, win it, and suddenly have 50 Seeding Points. The Big Three should be counted every year, whether you play them or not. And a few years should be counted, whether you’ve played in them or not. I don’t like punishing a new pair, but consistency is important, and veteran teams deserve some compensation for their experience.

Rate partnerships and teams, not just individuals

Another huge problem with the current system is that it rates individuals, not partnerships. As we all know, partnership is a huge aspect of the game. Let me use the top two pairs on the top-seeded teams from this year’sSpingoldas an example. As they lined up, the teams were:

Meckstroth-Rodwell; Levin-Weinstein

Helgemo-Helness;Fantoni-Nunes

Each of these players has 50 Seeding Points, so they would be seeded just the same if the teams were

Meckstroth-Fantoni; Levin-Helgemo

Rodwell-Helness;Weinstein-Nunes

Or some other random variation. I would expect these second teams to do pretty well, just based on the caliber of players, even in untested partnerships. But they would not be expected to be competitive against the first two teams.

Partnership matters. We shouldn’t just be rating players, we should be rating partnerships. So the Seeding Points need to be for a pair, not individuals. Results earned separately should be included in a pair’s total, but to a much lesser degree than the results earned as a pair. Obviously JeffMeckstrothdeserves to be rated highly, no matter whom he’s playing with. But when he plays withRodwellhis rating should go through the roof.

While we’re at it, teams should earn additional Seeding Points for results won together as a team. This shouldn’t be as dramatic as with pairs, but team chemistry is an important and often-overlooked factor. When you look at some of the great teams in history — the Blue Team, the Aces, theNICKELLteam — they were more than a combination of individuals or pairs; they were a team. MeckstrothandRodwellare a great pair. Playing withNickell-Freeman andHamman-Wolff /Hamman-Soloway,they were better. Teams should be encouraged to stay together and should be rewarded for their accomplishments.

Conclusion

The current seeding method gets the job done, but it could easily be much better. It’s not good for the team of younger experts that falls into the cycle of no seeding points —> low seed —> lose early —> earn no seeding points. And it’s no good for the top teams who have to face this unheralded team in the first or second round. Most importantly, it’s not good for bridge. It creates a barrier to entry at a time when we need to be courting new players rather than turning them away. The great thing about this game is that it’s open — show up, pay your money, everyone has a fair shot. We need to fight to keep it that way.

78 Comments
Getting Comments...
.

Bottom Home Top