Join Bridge Winners
All comments by Steve Willner
You are ignoring the author of this comment. Click to temporarily show the comment.
Even if you don't regard that as proof, it seems consistent with “anything can happen in a short match.”
June 19
You are ignoring the author of this comment. Click to temporarily show the comment.
I agree “expected outcome” would be better, but we have to interpret the words we've been given. There are plenty of other places where careful language could make the Laws clearer.
June 18
You are ignoring the author of this comment. Click to temporarily show the comment.
There is often a tradeoff between the best method of choosing a winner and the social aspects of a competition. The latter suggests letting many teams continue, while the former suggests uniform pruning. Organizers have to decide the relative importance for particular events. That's why different events have different CoC.
June 18
You are ignoring the author of this comment. Click to temporarily show the comment.
“We've always done it that way” is never a good argument against innovation.

On the merits, having three teams reach Day 2 to determine a single winner wouldn't be my first choice of format, but _with the new VP scales_, it isn't ridiculous. (When only the old ones existed, it was ridiculous in any serious event.)

If the 3-way format is used, there are valid questions as to carryover and whether one team should be dropped halfway. Those don't invalidate the format.
June 18
You are ignoring the author of this comment. Click to temporarily show the comment.
It depends on how you view the contest. If all you care about is deciding the first two teams – which is how I look at it – cut the field on day 1, and play a RR scored by VPs on day 2. KO is a terrible format for determining second place unless you are willing to have a playoff after the main final. (The USBC does that, but I doubt many GNT-B/C teams will want to.)

The field cuts on Days 1 and 2 should be approximately equal factors. Perfect is 8 to 4 to 2 or 18 to 6 to 2. If you don't get perfect numbers, you'll have to decide how to round off, but tending to round up probably makes sense for social reasons and to collect more entry fees.

Carryover should preferably be based on prior matches against the same team. The Bethe VP scales are set up to make that work out right, i.e., your expected VPs against another team after Day 2 are independent of whether you did or did not play them on Day 1. (The variance of the VPs is smaller if you played them on Day 1, though, because it's a longer match.) The alternative is to base carryovers on VPs earned on Day 1, which depends on how teams did against teams that did not qualify. That has a lot of luck in it, especially if Day 1 was Swiss and not RR, because some teams will have played against very weak teams and crushed them, while others drew only “middle of the pack” and good teams and won modestly. Anyway, there's room for varying opinions here, but there should be some carryover to discourage dumping on Day 1 and to increase the length of matches.
June 18
You are ignoring the author of this comment. Click to temporarily show the comment.
Heh. Proving, once again, that anything can happen in a short match. If three teams are within 0.1 IMP/bd of each other, playing 30 boards can't separate them reliably. However, the team in last place after 30 boards is unlikely to be (but still might be) the best team of the three.
June 18
You are ignoring the author of this comment. Click to temporarily show the comment.
Just to be clear, we're talking about IMP carryover from the prior play against one specific team, right? How did a team get to be “minus 100” in a 7 or 8 board match?

If a team is starting out -100, or more realistically -25 or so even on a bad day, why should all those lost IMPs come back? And what about the team that won those IMPs? Is their good play worth nothing? If a team doesn't “try their hardest” for whatever reason, why is that anyone's problem but their own?
June 18
You are ignoring the author of this comment. Click to temporarily show the comment.
“Probable outcome” is in scoring units (matchpoints or IMPs), not in raw scores. In practice, it's the weighted average of possible outcomes.
June 17
You are ignoring the author of this comment. Click to temporarily show the comment.
If you've already bought the boards, how about white inserts for NV? (My first thought was green, but that's not so good for the color-blind.)
June 17
You are ignoring the author of this comment. Click to temporarily show the comment.
I'd say also in David's case. The board was not in accordance with Law 2 – there was no unique vulnerability condition marked on the board – so it has to be thrown out. Play a substitute board if time allows.
June 17
You are ignoring the author of this comment. Click to temporarily show the comment.
The OP (and others) might want to take a look at how the new VP scales were derived. They are based on a couple of assumptions consistent with observed match data. Given those assumptions (and the max VP of the scale, which is arbitrary and won't affect any outcome, and also ignoring roundoff to 0.01 VP), no other VP scales are possible.
June 17
You are ignoring the author of this comment. Click to temporarily show the comment.
Why “no VP scoring?” If you must start with three teams and eliminate one, the new VP scales are a good way to do it. No method is perfect without a large number of boards, but it's unlikely the team that gets dropped is far better than both of the teams that remain. VPs seems better for the purpose than net IMPs because of the way the new VP scale is defined. (The old VP scale would be worse.)

The one objection I can see is potential unsportsmanlike dumping. If, say, Team A bribes Team B to dump to A but play hard against C, that would be bad. It's not a huge advantage for A, though. Once B goes home, which they presumably will after losing big to A, what's left is a straight KO between A and C. A could just as well have bribed B to pretend illness and go home without playing.

Another wrinkle is that carryovers should count, but they should be in the form of net IMPs in previous head-to-head matches, not VPs from the RR that includes teams that didn't qualify. In the example case, on Saturday 7 teams play a RR with 8 boards in each match. Sunday first session, the 3 top teams play a 3-board RR of, say, 15 boards. Each match is then scored as a 23-board match including the head-to-head Saturday results. Lowest VP score goes home, and the remaining teams play 30 boards head-to-head. Final victor is by net IMPs on a 53-board KO (8 boards Saturday, 15+30 Sunday).
June 17
You are ignoring the author of this comment. Click to temporarily show the comment.
If you want overtricks or small differences in game or slam contracts to be worth less, one way to do it would be to score in ratios rather than differences. I've been meaning to post something about how this would work, but it's not hard. You probably wouldn't want to score a pair game by hand, though.
June 14
You are ignoring the author of this comment. Click to temporarily show the comment.
“The formulas used for the WBF and ACBL NABC+ Continuous are identical, with both rounded to 0.01 as that is the smallest difference under ACBL regulations allowed to determine a margin of victory.”

Keith should know, but the Codification doesn't say anything about rounding. If you do that, you have to be careful to preserve the second derivative negative, as Andy explained. Does anyone know whether the ACBL and WBF roundoffs are identical?

A minimum margin of victory does not imply need for roundoff. In fact, it would be better not to round off if a minimum margin is required.

Even if there are slight flaws at the 0.01 VP level, the Bethe scales are far superior to the old scales.
June 13
You are ignoring the author of this comment. Click to temporarily show the comment.
“No one with under 2,000 masterpoints needs to know more than one defense to notrump or carding method! Let them focus on what really matters. We need to get the top teachers together and create a curriculum everyone can use. This will lower the barrier to entry tremendously.”

I agree with that last part. I learned from a Certified Goren Teacher, and in those days I could sit down opposite anyone and play with no discussion of methods. I relived those days for one session in the 1990's, playing with a first-time partner who had also learned Goren methods. (We had a 30-second discussion of changes from Goren.) SAYC was supposed to be a standard system everyone could use, but it's such a dog's breakfast that nobody wants to use it. Having something simple and reasonable would be very helpful.

I disagree completely with that first part. I have well under 2000 mp, but I bet I've played more systems (and more defenses to 1NT) than at least 95% of readers here. That's part of the fun. I might agree with “under 200 mp” and would probably agree with “under 20 mp.” I might be wrong even there, though. We have some new players in our club that came up with their own amusing defense to a strong 1 opening. They seem to like it, and nobody at the club (mostly under 1000 mp) has any problem with it, even though it would be illegal under the ACBL Convention Charts because it features a “Purely Destructive” overcall. (To be fair, only a few pairs at the club play strong club methods and are affected.)

Giving newcomers a standard framework is great. Forbidding them to do anything else is not a good idea.
June 13
You are ignoring the author of this comment. Click to temporarily show the comment.
The rules of Hool are slightly unclear in places, but there seems to be a rudimentary auction before play. The scoring system also differs.

It might make sense to introduce Hool as a step between Minibridge and real bridge, but having even a rudimentary auction makes no sense to me before the players understand trick-taking. I don't see why different scoring is an advantage. That said, Hool has a lot of deep aspects and could be quite a good game on its own.
June 13
You are ignoring the author of this comment. Click to temporarily show the comment.
Someone who knows and says what his biases are is preferable to someone who doesn't know or won't say.
June 13
You are ignoring the author of this comment. Click to temporarily show the comment.
Just be aware that for bridge, ‘contestant’ is specifically defined in the Laws. It means the whole team for team play, the pair for pairs play, and one person in an individual event.

Reading the definition of contestant makes it clear to me that ‘player’ means what John wrote, at least in that definition.
June 13
You are ignoring the author of this comment. Click to temporarily show the comment.
Putting this here because the sub-thread has moved on. When I wrote that the old VP scale was “obsolete,” I meant in terms of utility or wisdom of use. The old VP scale is legal to use if someone wants to. It has the merit of simplicity, but for a serious event, anyone with any sense would use the new (Bethe or continuous) scale.

I am pretty sure ACBLscore has the new scale programmed in. If not, it's easy to make a spreadsheet to calculate the VPs. According to the ACBL codification referenced above, VPs should be calculated to full floating point significance, not rounded to 0.01 VP. This differs from the WBF scale, which is rounded.
June 13
You are ignoring the author of this comment. Click to temporarily show the comment.
1. did the original CoC – the ones with two copies sitting on the Director's table – specify a VP scale? If not, who made the decision to use the obsolete scale instead of the current one?

2. as some others have asked, if “suggested improper player behavior” includes bridge actions, why can't adjusted scores affect the result?
June 10
.

Bottom Home Top