Join Bridge Winners
All comments by Mike Doecke
You are ignoring the author of this comment. Click to temporarily show the comment.
My experience is that:

X = GF, no clear direction.
Bids = Natural, F1 @ 1 Level, GF @ 2+ Level, 1NT = INV.

is very effective.

I actually prefer facing this overcall method because the opponents often can't preempts as effectively on the first round of the auction because they don't know what their partner has!
June 25, 2015
You are ignoring the author of this comment. Click to temporarily show the comment.
I really like the fact that your method allows the balanced hand to declare the NT when opener is strong and shapely. However, you haven't mentioned how responder deals with the G/F club hands. I think it's more valuable for responder to be able to show rather than ask on these hands.

Playing in the rest of the world where 1M - 2C can be 3-way (G/F Clubs, G/F Balanced or Limit Raise), I use a similar (but slightly less complex) method after 1H - 2C :

2D - Accepting a limit raise, no slam interest. Typically (11)13-16 OR 18/19 Balanced.

2H - Rejecting a limit raise.

2NT - 6+H, Extra values, but not a solid suit.

3H - Self supporting suit, usually balanced.

2S/3C/3D - Natural, slam interest opposite a limit raise.

After the auction starts: 1M - 2C - 2D responder can bid:

2H - 3H and slam interest. (either clubs or balanced)
2S - Clubs, without primary fit.
2NT - Balanced without 3H.
3H - 3H and no serious slam interest.
3D/3S - Natural, G/F with 6+ Clubs and serious slam interest.
4H - Limit raise.

Having played this method for a few years, I agree that the biggest gain comes from allowing opener to quantify their range immediately and unambiguously.

An unintended side-effect of including limit raises in 2C is that it helps divide up the range even more.

On a majority of hands this is enough information for responder to place the final contract and avoid giving away unnecessary information.






May 12, 2015
You are ignoring the author of this comment. Click to temporarily show the comment.
Explaining South's double as ‘primarily penalties’ feels like a misnomer to me also, especially given that North chose to bid on this hand. (And Kit didn't comment that it was an unusual choice)

It might only be a semantic difference but my agreement on this double would be “Extra values, not extreme shape”. The fact that partner will usually pass (almost always when holding a flat hand) is a function of the high level rather than the meaning of the double.

Kit: I'm genuinely curious as to how you would act given the same auction holding:

KQJx xxxx KQJ Ax
x AKxx AQxx KQxx
xxx AKxx AKx AKQ

ps. thanks again for an amazing column!
May 9, 2015
Mike Doecke edited this comment May 9, 2015
You are ignoring the author of this comment. Click to temporarily show the comment.
As soon as you start reducing the K value you run into the opposite problem of pairs who play just enough matches to maintain their ratings. Rating manipulation will always be an issue if every match counts.

Another reason that ELO doesn't translate well to bridge, is that chess players of significantly differing skill levels simply don't play against each other. In bridge a weaker team can (occasionally) win a set of boards against a stronger team which is an attractive feature of the game!

I'm not claiming that an ELO style rating system would be meaningless in bridge. If the ratings were calculated purely as a matter of interest, and were calculated independently on different subsets of tournaments (local, regional, national, elite) then they would provide an interesting opportunity to quantify the best players.
April 13, 2015
You are ignoring the author of this comment. Click to temporarily show the comment.
Yep - there are plenty of other ways to improve the simulation. I was just sticking with the criteria stated in the OP.
April 13, 2015
You are ignoring the author of this comment. Click to temporarily show the comment.
I don't have a strong opinion about 1H, but i'm suprised no-one has mentioned the possibility of a 1S advance. The fact that this hand competed to 3H over 3D anyways, suggests there seems little cost in showing spades along the way.
April 13, 2015
You are ignoring the author of this comment. Click to temporarily show the comment.
Creating an ELO style rating system for bridge is very feasible. It would even be possible to back-date such a rating based on results from many years of national/international events.

However, an ELO based system isn't without flaws. It offers an incentive for players to NOT play in order to protect their ratings for serious events. It would also be of limited benefit to the 90% of bridge players that don't play the game competitively. Bridge federation's gain far more with their current master-point based awards model.

Proof of this can be seen in the evolution of the “Magic the Gathering” rating system. In the early 2000s they switched AWAY from an ELO based system for event qualification and towards a participation based system, more in line with masterpoints. This was clearly a successful choice because their player based can grown consistently to more than 10 million players worldwide.


April 13, 2015
You are ignoring the author of this comment. Click to temporarily show the comment.
I was a bit skeptical of Ed's results so I ran my own simulation using the same assumptions. I've attached the .tcl code at the end (I used “DEAL” by Thomas Andrews).

I only considered the potential of the hands in NT.

Over the 10000 deals that I tested:

West playing in NT made:

7+ Tricks - 86.6%
8+ Tricks - 60.5%
9+ Tricks - 23.6%

Considering how partner would deal with the invite, I looked at only the most simple case: Pass with 15-16, Accept with 17-18.

Using Kit's suggestions about what deals matter:

Accept & Make 3NT = 11.5%
Accept & -1 in 3NT = 8.8%
Reject & Go down in 2NT = 23.2%

In summary, my analysis suggests that inviting in NT is a big loser in the long run, at least -0.5 IMPs/board.

—– .tcl code below —–

east is JT87 AT5 8643 Q6

if {diamonds south < 4 or spades south > 4 or hearts south > 4 } {reject}

if {hcp south < 12} {reject}
if {hcp west < 15 or hcp west > 18} {reject}
if {diamonds west <3 } {reject}
reject if {not semibalanced west]}

set tricks dds west notrumps

if { < 17} {puts “Reject $tricks”}
if { > 16} {puts “Accept $tricks”}

accept
April 13, 2015
Mike Doecke edited this comment April 13, 2015
You are ignoring the author of this comment. Click to temporarily show the comment.
Another great article Sartaj. Although i'm surprised you didn't lead a trump!
Dec. 1, 2014
You are ignoring the author of this comment. Click to temporarily show the comment.
I prefer 1S rather than 2S because it saves a level of space and avoids the ambiguous continuations after a SJS.

Having made a SJS, I strongly prefer 3H next which is economical and keeps that strain in the picture.

Over 5C I'll guess to pass.


Nov. 25, 2014
You are ignoring the author of this comment. Click to temporarily show the comment.
Another great article.

However your next article should be “Transfers after 1C: the best of both worlds!”

March 7, 2014
You are ignoring the author of this comment. Click to temporarily show the comment.
Gierulski didn't have a problem, the H9 is a blunder.
May 4, 2013
You are ignoring the author of this comment. Click to temporarily show the comment.
Excellent analysis!
Feb. 8, 2013
You are ignoring the author of this comment. Click to temporarily show the comment.
Great summary. I've provided a few alternative treatments below.

I prefer to use 1M 2C 2D as an artificial relay on any hand without significant extra values (roughly 12-15 HCP). This limits openers hand immediately and avoids consuming bidding space that responder can better use to clarify their hand type at the 2 level. If opener does have extra values, they can show them immediately by reversing or responding naturally at the 3 level.

Playing your suggested method, the auction might start 1M 2C 3C 4C and the partnership has reached the 4 level with neither player showing or denying extra values. Even responder's hand type still isn't clear. Playing the artificial 2D response, the auction might start 1M 2C* 2D* 2NT 3C and now responder holding a minimum balanced hand with club support, might choose to investigate 3NT rather than committing to 5C.

Playing the 2D artificial relay, you might also choose to add limit raises into the 2C response. In this case 1M 2C 2D 2M shows a limit raise and allows the partnership to stop in 2M if partner rejects the invite. Using this treatment makes responding on strong hands with 3M fit more awkward, but it might be a worthwhile trade-off especially in the context of a light opening, strong club system.

Another treatment that is useful for partnerships who prefer to open very aggressively in the majors uses 1M 2C 2M as an artificial rebid, showing a sub-minimum opening (10-11 HCP) but not promising extra length. This treatment comes up remarkably often and allows the partnership to stop in 2M on hands where responder might otherwise have forced to game and to stop in game on hands that might otherwise have tried for slam. Hands that would usually rebid 2M can either be dealt with using a 2M/2NT inversion, or simply left in the 2D relay.
Dec. 20, 2012
Mike Doecke edited this comment Dec. 20, 2012
You are ignoring the author of this comment. Click to temporarily show the comment.
Nice hands!
Nov. 28, 2012
You are ignoring the author of this comment. Click to temporarily show the comment.
Another great article.

One hidden side-effect of making your overcalls as destructive as this is that it reduces your partner's ability to further the preempt. Another consideration is that partner already had a chance to open a destructive preempt in first seat and didn't, so perhaps that also should be a mild dissuading factor.
March 28, 2012
.

Bottom Home Top