[irdoz]

Prob best to give link...

http://www.migry.com/Articles%20and%20othe...he%20winner.pdf

The article gives no. of hands analysed. It doesnt do frequency that different ranges occur (lots of other places do) but it does give a frequency table comparing 1992 to 2002 of the different ranges pairs are actually playing.

[effervesce]

matmat, on Mar 18 2008, 08:03 PM, said:

irdoz, on Mar 18 2008, 06:58 PM, said:

I came across this data (published in an isreali bridge magazine) based on an analysis of hands at european/world championship level between 1992 and 2002.

Average imp score when you open 1nt

Overall +0.52 imps average
16-18 -0.21
15-17 0.57
14-16 0.24
13-15 0.43
12-14 0.53
10-12 0.69

(I think to make a real comparison of the effectiveness of different ranges youd have to include 2nt data for hands opened 1minor etc.)

Heres the average if you 'stretch' and open a 1nt 1+ points below yr stated range...

Overall -0.31
15-17 -0.28
14-16 -0.76
13-15 -0.44
12-14 -0.56
10-12 +0.74

Hmm... seems at this level on average stretching doesn't work out well (with 10-12 being a big outlier)

i find it curious that there is a peak at the 15-17 range. there's a clear trend in increasing imps from the higher ranges to the lower ranges otherwise.

any idea why?

I dont know, maybe its that the weaker 'strong NT' preempts the opps more effectively when it is opened compared to 'weak NT', and thus they're still as scared to overcall a 15-17NT just like a 16-18 as it is still a 'strong NT'? Whereas for 14-16 people tend to overcall more, treating it more often as a 'weak NT'?

I'd also wonder about the effect of 'variable NT' openings- if people open one range in 1st/2nd or 3rd/4th, or one range vul and another non-vul, that perhaps may have influenced the IMP gain. 10-12 tends to be opened in 1st/2nd seat - perhaps the preemptive value is greatest in such seats, while retaining constructive bidding with partner. Such 'seat effects' may perhaps be why 14-16 and 13-15 seem less effective.

This idea is supported by the fact that the 10-12 NT is the ONLY range that gains IMPs when opening light - it appears that the main gain of 10-12 NT is its preemptive effect, which more than counterbalances the loss in constructive bidding. Other NT ranges lose more in opening light from constructive bidding than gain in preemptive bidding.

[matmat]

irdoz, on Mar 18 2008, 08:12 PM, said:

Prob best to give link...

http://www.migry.com/Articles%20and%20othe...he%20winner.pdf

The article gives no. of hands analysed. It doesnt do frequency that different ranges occur (lots of other places do) but it does give a frequency table comparing 1992 to 2002 of the different ranges pairs are actually playing.

hmm... i obviously failed reading comprehension... does this explain why 15-17 is such a big winner?

[effervesce]

matmat, on Mar 18 2008, 08:58 PM, said:

irdoz, on Mar 18 2008, 08:12 PM, said:

Prob best to give link...

http://www.migry.com/Articles%20and%20othe...he%20winner.pdf

The article gives no. of hands analysed. It doesnt do frequency that different ranges occur (lots of other places do) but it does give a frequency table comparing 1992 to 2002 of the different ranges pairs are actually playing.

hmm... i obviously failed reading comprehension... does this explain why 15-17 is such a big winner?

No it does not - it only explains why more and more people are turning from 16-18 to 15-17 - that the disadvantage of the 1x-1NT rebid being 12-15 outweighs the benefit of a precisely 19 pt 1x-2NT rebid in a 16-18 pt range.

However, this does NOT explain why an opening bid of 15-17 1NT performs better than 16-18.

[han]

I really don't know what conclusions to draw from this without knowing how many times each notrump range was opened. This is especially true for the upgrades.

[jtfanclub]

effervesce, on Mar 18 2008, 10:33 PM, said:

No it does not - it only explains why more and more people are turning from 16-18 to 15-17 - that the disadvantage of the 1x-1NT rebid being 12-15 outweighs the benefit of a precisely 19 pt 1x-2NT rebid in a 16-18 pt range.

However, this does NOT explain why an opening bid of 15-17 1NT performs better than 16-18.

It doesn't even explain that...how many of the best players play a 16-18 NT? How does their quality of play compare to those who play 15-17?

I also don't think it's fair to compare all seats equally. People who play a 10-12 1st and 2nd and 15-17 3rd and 4th may have an advantage because their partners are more restricted than playing 16-18 or 15-17 in all seats.

[effervesce]

jtfanclub, on Mar 18 2008, 11:09 PM, said:

effervesce, on Mar 18 2008, 10:33 PM, said:

No it does not - it only explains why more and more people are turning from 16-18 to 15-17 - that the disadvantage of the 1x-1NT rebid being 12-15 outweighs the benefit of a precisely 19 pt 1x-2NT rebid in a 16-18 pt range.

However, this does NOT explain why an opening bid of 15-17 1NT performs better than 16-18.

It doesn't even explain that...how many of the best players play a 16-18 NT? How does their quality of play compare to those who play 15-17?

I also don't think it's fair to compare all seats equally. People who play a 10-12 1st and 2nd and 15-17 3rd and 4th may have an advantage because their partners are more restricted than playing 16-18 or 15-17 in all seats.

That raises an interesting point - perhaps another advantage of the 1NT 10-12 opening is the probability of playing in 1NT. As 1NT is probably THE MOST DIFFICULT contract to defend against, the gain for the 10-12 1NT could be due to 'declarer advantage'. This advantage is probably much smaller in a contract of 3NT, which is probably easier to defend than 1NT.

[matmat]

effervesce, on Mar 18 2008, 11:16 PM, said:

That raises an interesting point - perhaps the greatest advantage of the 1NT 10-12 opening is the probability of playing in 1NT. As 1NT is probably THE MOST DIFFICULT contract to defend against, the gain for the 10-12 1NT could be due to 'declarer advantage'. This advantage is probably much smaller in a contract of 3NT, which is probably easier to defend than 1NT.

I am also not really sure how valid it is to extract the NT range from the context of the remainder of the system. 1nt 10-12 on hands where you get to open it might well gain you half an imp a board, but on deals where you have to use a different sequence to show the balanced point counts that others open 1nt on, you might be losing that same half an imp a board, or some such.

[effervesce]

matmat, on Mar 18 2008, 11:18 PM, said:

effervesce, on Mar 18 2008, 11:16 PM, said:

That raises an interesting point - perhaps the greatest advantage of the 1NT 10-12 opening is the probability of playing in 1NT. As 1NT is probably THE MOST DIFFICULT contract to defend against, the gain for the 10-12 1NT could be due to 'declarer advantage'. This advantage is probably much smaller in a contract of 3NT, which is probably easier to defend than 1NT.

I am also not really sure how valid it is to extract the NT range from the context of the remainder of the system. 1nt 10-12 on hands where you get to open it might well gain you half an imp a board, but on deals where you have to use a different sequence to show the balanced point counts that others open 1nt on, you might be losing that same half an imp a board, or some such.

I meant that the gain comes from playing a contract of 1NT, as opposed to 3NT. An opening bid of a 10-12 1NT is more likely to play in 1NT as opposed to 3NT. An opening bid of 16-18 is less likely to play in 1NT as opposed to 3NT.

This explanation, whereby the increased likelyhood of playing in 1NT as opposed to 3NT when the point range is reduced, may explain the negative correlation between 1NT strength and IMP/board gain.

For example, an opening bid of 10-12 NT (just a number for example) may have a 50% chance of playing in 1NT, and a 15% chance of playing in 3NT. An opening bid of 16-18 1NT may have a 50% chance of playing in 3NT, and 15% chance of 1NT.

If, whenever you play in 1NT the gain is 1 imp, and the gain in 3NT is -0.5 IMP, then we see an advantage of opening a 10-12 1NT as 0.5x1 - 0.15x0.5 = 0.425 imp average every time it is opened, while for the 16-18 1NT would be 0.5x-0.5 + 0.15x1 = 0.1 imp every time a 16-18 1NT is opened.

[matmat]

effervesce, on Mar 18 2008, 11:30 PM, said:

If, whenever you play in 1NT the gain is 1 imp, and the gain in 3NT is -0.5 IMP, then we see an advantage of opening a 10-12 1NT as 0.5x1 - 0.15x0.5 = 0.425 imp average every time it is opened, while for the 16-18 1NT would be 0.5x-0.5 + 0.15x1 = 0.1 imp every time a 16-18 1NT is opened.

okay...

and what about the loss from defending 2M and either not setting it or setting it not enough when you open 1m instead of 1nt?

[effervesce]

matmat, on Mar 18 2008, 11:39 PM, said:

effervesce, on Mar 18 2008, 11:30 PM, said:

If, whenever you play in 1NT the gain is 1 imp, and the gain in 3NT is -0.5 IMP, then we see an advantage of opening a 10-12 1NT as 0.5x1 - 0.15x0.5 = 0.425 imp average every time it is opened, while for the 16-18 1NT would be 0.5x-0.5 + 0.15x1 = 0.1 imp every time a 16-18 1NT is opened.

okay...

and what about the loss from defending 2M and either not setting it or setting it not enough when you open 1m instead of 1nt?

That was one of the reasons I suggested earlier for the lower gains for opening 14-16 or 13-15. You have less of the 'dont overcall this' scare-factor for a 14-16 or 13-15 opening as compared to a strong no-trump. Maybe the gain in 10-12 1NT is 'I cant overcall this without constructive values'.

I dont know there are lots of variables to consider.

Anyway, your point about the loss of opening 1m due to opps playing 2M is not as relevant to this study-its about the gain from making a 1NT opening bid. Yes, these gains may be offset elsewhere. What we are trying to work out here is WHY certain point ranges have a greater +IMP/board score when opened than others.
In any case, for any range, there will be different hands that the opps can make an overcall on when you open 1m. It'd be hard to compare and quantify such results.

[benlessard]

Quote

I've seen it only once. The pair in question had a 13 - 16 rebid of 1NT. Ugh...

Romex use 1Nt openings for the unbalanced 18-22 hands. So the 1Nt rebid is quite wide. Im not saying i like it but its playable.

We play wide ranging Nt (10-14) & 12-15 and wide ranging rebid after a strong 1C (15-18 or 16-19) Im convinced that with a proper structure the upside can be higher then the downside.

Where we are winning the most is

when NV the 10-14 range is a killer.

1D---1M----1Nt (is showing 6 diamonds)
1H---1S-----1Nt (is showing clubs)
etc.

Where we are losing the most is when

1C---1D----1Nt (16-19) and responder with 7 pts make an invitation.

[matmat]

effervesce, on Mar 18 2008, 11:49 PM, said:

Anyway, your point about the loss of opening 1m due to opps playing 2M is not as relevant to this study-its about the gain from making a 1NT opening bid. Yes, these gains may be offset elsewhere. What we are trying to work out here is WHY certain point ranges have a greater +IMP/board score when opened than others.

And it is kinda stupid to try to quantify the effect of opening 1nt on different ranges the way it is done here. you have no information about the quality of the players, you have no information about the rest of their system.

When you make a decision to switch from 16-18, to, say, 10-12, just because of this study, you may gain close to an imp every time you open 1nt, but is it really worth it if the rest of your system losers a third an imp per deal?

i feel like i'm repeating myself.

[effervesce]

matmat, on Mar 19 2008, 12:36 AM, said:

effervesce, on Mar 18 2008, 11:49 PM, said:

Anyway, your point about the loss of opening 1m due to opps playing 2M is not as relevant to this study-its about the gain from making a 1NT opening bid. Yes, these gains may be offset elsewhere. What we are trying to work out here is WHY certain point ranges have a greater +IMP/board score when opened than others.

And it is kinda stupid to try to quantify the effect of opening 1nt on different ranges the way it is done here. you have no information about the quality of the players, you have no information about the rest of their system.

When you make a decision to switch from 16-18, to, say, 10-12, just because of this study, you may gain close to an imp every time you open 1nt, but is it really worth it if the rest of your system losers a third an imp per deal?

i feel like i'm repeating myself.

Yes, I get your point, you get mine. As for your point about the rest of the system losing out-yes that may be true in a standard system, but some systems (eg strong club) actually have it the other way around, benefiting from the opening of a 10-12 NT - a synergistic effect on the rest of the system whereby using such a range means that other openings are more clarified. For example, putting all 1st/2nd seat 10-12 5M332 in 1NT means that 1H/1S are either unbalanced or maximum in a precision context. It makes inviting opener's 1M much less risky if opener is unbalanced or maximum than either unbalanced and minimum flat.

It also allows a 3rd eat preempter more freedom to open 1M, or make a 3x/4x preempt given the reduced chance of game/slams respectively.

[awm]

In addition, these things are probably vulnerability-sensitive. Few if any top-flight players use a notrump range of 10-12 at vulnerable. It wouldn't surprise me if opening notrump in general was more of a winner at non-vulnerable because most of the losses involve playing in the wrong partial, which is never a big minus position at non-vulnerable (you can lose a lot of imps for -200 or -300, not that many for -100 or -150 especially since the opponents probably rate to go plus in a partial when you are getting these scores).

I suspect that any reasonable attempt to analyze 10-12 notrump in third seat vulnerable would indicate that it's a terrible idea. The thing is that no top flight pair plays this, so there's not really any data on it.

[skjaeran]

I don't understand some arguments here about 15-17NT getting less overcalls than weaker ranges. To me it's just the opposite - there's a lot of hands I hold where I regularly overcall stong 1NT (14-16 or stronger) where I'd be very reluctant to overcall a 13-15NT and never consider a bid over a weaker range.

You have to get in over opps strong 1NT as often as possible to disrupt their constructive sequences on hands you can't bid vs a weak NT, where you've got to cater for your own constructive issues.