Join Bridge Winners
All comments by Michael Bodell
1 2 3 4 5 6 7 8 9 ... 151 152 153 154
You are ignoring the author of this comment. Click to temporarily show the comment.
An easy evaluation of quality of stratification if you have one measure of quality used for strats and a different one for quality of strat would be how many people are in the right start if you were using the other metric for strat. I.e., if you use master points for the strats and have A B and C pairs that is one measure. Then if you use something like power ratings that is a different measure and you can see how many people would change strats if you switched to doing the strats by power rating.

You could also measure quality of strat by looking at the results and seeing do the A players beat the B players who beat the C players. If your strats are good this will tend to happen. This doesn't work for flights, but does for strats. The other check you can do is inside the strat do the players finish in order of your measure. That is if you strat by MP, then do the flight A players with more MP do better than the flight A players with fewer MP. Likewise for other strats. If your measure is power ratings, you can similar compare results to expected results. This sort of comparison (when done with enough data) can also show is something like master points predictive of results? If so, is it more or less predictive than alternative measures like power ratings. (The answer is MP have some predictive value but PR have more predictive value).
Feb. 20
You are ignoring the author of this comment. Click to temporarily show the comment.
I wouldn't include NT range in 2/1 GF as a label on its own.

If I sat down in the US and said “2/1 GF” it is true that I expect that we'd play a strong NT, but that isn't because “2/1 GF” implies a strong NT, it is because I sat down in the US. I'd also expect a strong NT if I sat down in the US and said “Hi I'm Michael” or “standard american” or “udca” or whatever as my only exchange with partner. But surely “udca” doesn't imply strong NT even if I'd assume we were playing it if our only verbalized agreements were “udca”.

If my partner and I play 2/1 GF and also a weak NT and I was describing the system to someone I would not feel I misdescribed it if I said we play 2/1 GF. That would still be what I'd put in the “general approach” line of the convention card. The same would be true if we played mini-nt or variable nt ranges.

I think it will be fine if Bridge Winners 2/1 has one defined NT range that is strong (and also fine if it allows leafs with different NT ranges - but that seems outside the scope of the current project), but I don't think “2/1 GF” has a fixed NT range attached definitionally.
Feb. 20
You are ignoring the author of this comment. Click to temporarily show the comment.
Me too. 4 cards in the major, any strength, at least 3 diamonds if 1 with 4=4=3=2 is allowed, 4+ otherwise.
Feb. 16
You are ignoring the author of this comment. Click to temporarily show the comment.
My comment was agreeing with Chris's above, not commenting either way on yours (sorry there isn't more elaborate threading in conversation).
Feb. 15
You are ignoring the author of this comment. Click to temporarily show the comment.
Yeah the 6 and 5M clearly need to bid the major over 2.
Feb. 15
You are ignoring the author of this comment. Click to temporarily show the comment.
If I were forced to guess if partner had explained my auction right and then bid 5 it is either a hand that has 6 diamonds and 1 heart but tried 2NT (3=K=6=3 or the like) or was a hand that had good heart support and diamond concentration looking for maybe slam? 2=4=5=2 Sort of a fit jump I guess.

Neither fits that well, since 4 and slower routes were available.

There are also the mis-sorting bids like all my red cards were diamonds and I have no hearts (pass wins) or all my diamonds were hearts and I invented an exclusion call (changing partners wins).

So I think maybe pass and 5 are possible bids if both above hands are possibility since it depending on what we think partner has. I'd pretty much certainly bid 5 with no UI, since I'm not really sold on pass overall. With the UI passing is much more likely to be wrong (since partner thinks I have diamonds and I don't), so if pass were an LA, it is certainly the one I must select. But I suspect there will be disagreement on if pass is really an LA.
Feb. 15
You are ignoring the author of this comment. Click to temporarily show the comment.
I'm fine with either of the first two, or even a blend where there is judgement (I.e., 2 shows serious interest in diamonds, 2M a 4-card major, but 2 doesn't deny a 4cM and 2M doesn't deny 5 mediocre or worse diamonds).
Feb. 15
You are ignoring the author of this comment. Click to temporarily show the comment.
Except aren't the videos being made public for multiple reasons, one of which is the enjoyment of watching them. The whole training, validation, test setup makes sense from the POV of catching cheaters, and makes sense if catching cheaters is the only reason for the videos, but part of the reason ought to be to be to share the spectating as well.

I mean you could make the same argument about training, validation, and test for the BBO vugraph hands as well, but again catching cheaters is only one of many purposes for vugraph - enjoyment of viewing is part of the goal too.
Feb. 14
You are ignoring the author of this comment. Click to temporarily show the comment.
A few quick comments:

1. It is really nice to see the statistics shared. There could likely be more slices or more data shared, but the perfect should not be the enemy of the good, and sharing this data is a definite good.

2. The 3300 players who played in this is a healthy number, a bit ahead of the ~2600 who tried the online individual experimental event. 3300 is also ~2% of our active ACBL member base, which is quite healthy especially given not everyone would be interested and not everyone would live near a club participating.

3. A couple of other questions: What percentage of the people who played a REACH eligible game had not registered? What percentage of those people would have won gold if they had registered and instead just got their club result? What percentage of people registered but then didn't show up in any club game (I.e., planned to play, but didn't)?
Feb. 13
You are ignoring the author of this comment. Click to temporarily show the comment.
Nice article.

I think another under appreciated reason is that older players have already a thought of what the agreed “standard” meaning of many bids and bidding sequences are, and hence think the “conventions” are special requiring more thought, memory, work, etc. Where a newer player (which younger people tend to be) may not know the standard meaning of the bidding sequence and hence approach it from a fresh perspective. For them, any meaning, “standard” or “conventional” or what not is equally strange or equally valid, so there isn't really a clear cost to the convention, or at least not the same cost that there is for the more experienced player. And thus for them they can start with the question “What is the best meaning for this sequence?” rather than “What is the standard meaning of this sequence?”.

This can be emphasized also as “standard” drifts over time, so if a more experienced player learned a bidding sequence as meaning one thing 30 years ago, but if the new “standard” meaning of that same sequence is something else, from the perspective of the older person this new thing is “convention” and memory cost to learn. But to the younger newer learning the newer “convention” is the standard and only way to play it, and so not a cost.

You can see an example of the last from 3014 versus 1430 keycard. For some folks who learned it early 3014 is the “natural” keycard and they find it a memory load to play 1430. Whereas for many people who have learned more recently 1430 is the “natural” keycard and it is a memory load to play 3014.
Feb. 7
You are ignoring the author of this comment. Click to temporarily show the comment.
I agree, but I'd also add Hearthstone - which is sort of like the simplified online version of MTG - but also super popular and with young professional players and non-trivial real money prizes (and globally popular with Europe, China/Asia, and the Americas all having active populations of players).
Feb. 6
You are ignoring the author of this comment. Click to temporarily show the comment.
2NT then 3NT could show this. I think you generally want a stop (or half stop) but if you have a full 14 count outside hearts, partner needs something for their X, maybe there something includes a heart stop (or half stop), or maybe you get lucky and opponents try an outside suit, or maybe partner with nothing in hearts will scramble.
Feb. 6
You are ignoring the author of this comment. Click to temporarily show the comment.
Notice in this day and age young people playing a card game? Almost like there is a good demographic available there with natural cross over appeal to bridge.
Feb. 6
You are ignoring the author of this comment. Click to temporarily show the comment.
If you want consistency you can play that fast denies (in context) and slow shows (in context) so that the slower treatment might show hearts well stopped while the fast treatment might show hearts only partially stopped. Say 0.75-1.5 stops for direct and 1.5-2.5 for slow?
Feb. 5
You are ignoring the author of this comment. Click to temporarily show the comment.
@Andy, bid 2nt if your agreement is that's the default bid and raises show 4+ and rebids of opening show 6+ and 3-level new suits show extra values or shape, and thus 2nt default is all that is left. Partner doesn't need to raise to 3nt if they have side suits or others so if partner has 4 they'll bid 3 and we'll find the heart flaw. Similarly, if partner has delayed support or extra diamond length, they can find it. And most likely, when partner bids 3 over this (likely showing club concern with hearts stopped), we can bid 3nt now. 2nt doesn't end the auction.

Am I claiming this is the super best fantastic system ideally suited for this exact hand combination? Not necessarily. Am I suggesting this can be played easily and effectively enough for the no discussion mostly natural system being developed here? Absolutely. And the benefits on other hands are helpful as well.
Jan. 31
You are ignoring the author of this comment. Click to temporarily show the comment.
You can square that circle by deeming that having a trump beyond what you promise is more than a minimum. Thus a 4th trump when you only show/promise 3 is more than a minimum. But a 4th trump when you show/promise 4 is just a minimum.
Jan. 30
You are ignoring the author of this comment. Click to temporarily show the comment.
That hand is obviously a constructed extreme, but even there, what are some sample hands for partner to make a GF 2/1 response when you have 12 of the 20 hcp in the black suits. How many points must partner have in the red suits? How many points is it likely that partner has in the red suits? How likely are we wide open in one of those suits (especially if partner just raises to 3NT instead of bidding a suit)?

Even in this extreme case the downside of bidding 2NT doesn't seem overwhelming to me, and agreeing that your raise shows 4+ cards seems a good benefit.
Jan. 29
You are ignoring the author of this comment. Click to temporarily show the comment.
The closest I could find to the participating schools is this map with pins: https://batchgeo.com/map/00522d5068605bb99f41a06f4e24c249 linked off http://www.acbl.org/juniors/collegiate-initiative/ at “2017-18 Participating colleges/universities”. If I counted right there are 33 pins, but I didn't see an easy list.
Jan. 29
You are ignoring the author of this comment. Click to temporarily show the comment.
Yes. Traditionally it is 4415 - 1 shapes (3415, 4315, 4405, and 4414).
Jan. 28
You are ignoring the author of this comment. Click to temporarily show the comment.
@Paul, it doesn't change quite as much as you'd think. If you assume in both models your side has 26 hcp opposite the minimum threshhold and that you have nothing in the suspicious suit (I.e., you have 11 opposite 15-17 or 14 opposite 12-14 and you have nothing in hearts) then you can think of splitting the heart and non-heart points.

In the strong NT hand there are 10 heart points, of which you have none, and 30 non-heart points of which you have 11 so 19 non-heart points out there. Your strong NT partner has 15-17 of the 10+19 points, so likely about 1/3 (10/29) of their points, or about 5, are in hearts.

In the weak NT hand there are 10 heart points, of which you have none, and 30 non-heart points out there of which you have 14 so 16 non-heart points out there. Your weak NT partner has 12-14 of the 10+16 points, so likely about 40% (10/26) of their points, or about 5, are in hearts.

It isn't exactly the same if you use the more exact numbers, but it is pretty close to the same and the same system can be used.

I have used the same system even over mini-nt and it works well enough.
Jan. 26
1 2 3 4 5 6 7 8 9 ... 151 152 153 154
.

Bottom Home Top