Join Bridge Winners

Bridge Winners Profile for Martin Lindfors

Martin Lindfors
Martin Lindfors
  • 26
    Following
  • 1
    Followers
  • 17
    Posts
  • 0
    Favorites

Basic Information

Member Since
Sept. 24, 2017
Last Seen
39 minutes ago
Member Type
Bridge Player
about me

Amateur player in Linköping, Sweden.

Country
Sweden

Bridge Information

Member of Bridge Club(s)
Filbyter Bridge
ACBL Ranking
None
Sorry, this user has no cards yet.
Shailesh Gupta's bidding problem: KQTxx Jxx JT9xx ---
Yep. With 4 card support (change a red card) I would splinter (4).
Eric Dong's bidding problem: 863 KT742 QJ42 A
I don't play drury, but I do play inv+ Stenberg 2NT. Perhaps drury would be good on some hands, but not this one, I think. I happy to play in 3 after a quite subminimum 1 opening.
Equivalent of FSF-1round, NMF/Checkback after 1m-1h; 2m-?
How does opener respond after sequence (1)?
A Tale of Two Bootstraps OR How to significantly improve ACBL masterpoint allocations in a way that most everyone will hate…
I think you misunderstood my reply. I quite understand what you meant. I was replying to: "I'd bootstrap each round separately and add the results of each round to simulate the match..." As stated above, I think that this procedure is biased when results are correlated between rounds. Within ...
A Tale of Two Bootstraps OR How to significantly improve ACBL masterpoint allocations in a way that most everyone will hate…
Richard, if you're bootstrapping each round separately and adding the results, you're basically still assuming that the rounds are independent. I'm not sure how you would be doing the adding, but if round results are positively correlated, it is clear that averaging win percentages would overestimate the ...
A Tale of Two Bootstraps OR How to significantly improve ACBL masterpoint allocations in a way that most everyone will hate…
The moments of the data, yes. Not the true distribution.
A Tale of Two Bootstraps OR How to significantly improve ACBL masterpoint allocations in a way that most everyone will hate…
Richard, bootstrap preserves the moments of the data, not the true distribution. When the true distribution is not sufficiently well approximated by bootstrapping the data, the moments will be poor estimates.
A Tale of Two Bootstraps OR How to significantly improve ACBL masterpoint allocations in a way that most everyone will hate…
I like this idea. Presumably we would have something like this: expectedImpsPerBoard = f(ratingHome, ratingVisit) ratingChange = g(actualImps, expectedImpsPerBoard, numBoards) ratingHome = ratingHome + ratingChange ratingVisit = ratingVisit - ratingChange However, I think it might feel demoralizing for a strong team to lose ratings against a weak team just because they won by 20 ...
A Tale of Two Bootstraps OR How to significantly improve ACBL masterpoint allocations in a way that most everyone will hate…
How is this a statistical test? I see no testing at all going on.
A Tale of Two Bootstraps OR How to significantly improve ACBL masterpoint allocations in a way that most everyone will hate…
Bernard, you're absolutely right. That's why it would be patently stupid to use bootstrap methods on 4 board matches. With 60 boards to sample from, one would be hoping that this evens out. And it might, or it might not. I haven't done the calculations and simulations ...
.

Bottom Home Top