[AyunuS]

I'm a person that's never liked blackwood. I feel that aces or even keycards do not make slams. They are worth something, yes, but I've never felt it to be all that accurate of a description of the slam capability of a hand. So, I have created this new little point system of:
A: 5 points
K of a suit with a known fit: 4 points
Any other K: 1 point
Q of a suit with a known fit: 2 points

And then I have created this lovely little system to show how many such points you have:
5♣: 0-3, 10-11, or 18-19
5♦: 4-5, 12-13, or 20-21
5♥: 6-7, 14-15, or 22-23
5♠: 8-9, 16-17, or 24+

The point of this is that you can likely figure out which range they are in, given that you reached this point in the bidding correctly, and you get a more accurate description of how useful the hand really is at making a slam. You can usually figure out exactly how many keycards they have, since you should have some of the other good cards so there are only so many possibilities left for what they can have. Let's take a look at an example of how powerful this can be in action, on a randomly generated hand, rerolled until it looked like it'd stand a chance at a slam:

---

So, what's so great about this 5♦? Well, it contains a good bit more information than your standard blackwood or even rkcb. Here's what you can come up with from it. And you also already know they have A♦ from their cuebid earlier.
1. Your partner has 12-13 such points. You have enough that they can't have 18, and if they did, they probably would've bid higher earlier. And if they've shown even one control they can't be in the 0-3 range.
2. So if they have 12-13 such points, they must have exactly 2 keycards. Nothing else they could possibly have would add up to that. So it's either 5♠ or 6♠.
3. So what do you put them at? Assuming 12 points for simplicity, they could have any of the following:
♠AQ ♦A
♠KQ ♥K ♦A
♠A ♥K ♦AK
♠KQ ♦AK

Now just figure out if more than half of those will make a slam or not. I like how this looks since I consider all 4 of these to be pretty close to equally good at making a slam, meaning my points are seeming pretty accurate.
First one I'd say is a high chance of a slam, since if either the K♠ or K♥ is onsides, you should be able to make it.
Second one I'd say you have it easy since your hearts will run and so you likely only have the one loss in spades.
Third one well unfortunately, I'd say that you probably won't make it. With only ♠A, missing both ♠KQ, you'd likely only make it on a 2-2 trump split.
The fourth one you can probably get away with just the one loss in spades, but in any case, I'll give it more than 50%.

So more than half chance of getting that slam, so go for it! And they might even have 13 and then you get an even better chance. But in any case, I feel this does a very good job of splitting the line between where it's worth bidding a slam and where it isn't. Notice that if your partner only showed 10-11, then it's not worth going for a slam since too many of the possibilities would have too bad of a spade suit. I wouldn't bid a slam if my partner only had A♠ or (yikes) only K♠, as you'd likely need a 2-2 split for it to possibly work.

But is this really better than blackwood or rkcb? I say yes, since you know what would happen if south used one of those? He'd be like, oh no, we're missing an A, and I don't have a good spade suit, so I guess I'll just not bid a slam. It doesn't get you extra goodies like guaranteeing some extra kings in the event that the ♠A is missing so it's really hard to be as confident. Like in my 2nd and 4th cases, if you don't know that you have ♥K or ♦K, you probably decide not to bid the slam. But those are crucial since they mean you likely only have losers in the trump suit, which means that you likely only have the one loser.

Now, I know what the counterargument to all this would be. They'd simply say that rkcb would still find slam. How? After the 5♦, a lot of good players bid 5♥ to ask for the ♥K. Then partner can show it and then you can safely get to 6♠ even without all this extra stuff. While this is true, it took it somewhat more bidding space to reach a comparable amount of information. What I mean is that, if I already know more at 5♦, I could use 5♥ for some other meaning to get even more information to make an even better decision than the rkcb player would make. Additionally, the rkcb player would have to bid ♠ over 4NT so they wouldn't even have room to check for anything like ♥K.

One could also argue that it's too each to get stuck at 5♠ with this system with hearts trump. There's an easy solution to this-make it so that 4♠ means this and shift all the responses down by 1. There's no rule that says only 4NT can be this kind of bid.

Anyway, let me know what you think. Is this a good idea? Is there some way to make it better? (Well duh, it doesn't even count singletons or voids). If it's bad, show me a hand that it does a bad job at dealing with, so I can get an idea as to how to maybe fix it.

[mgoetze]

AyunuS, on 2016-January-18, 22:56, said:

And you also already know they have A♦ from their cuebid earlier.

Aha, so your system requires playing cuebids strictly as first-round controls. I stopped reading at this point - there are very good reasons why basically the whole world has moved on from first-round control cuebids to mixed cuebids, if you want me to read the rest you'll first have to convince me that I don't need mixed cuebids.

[nullve]

mgoetze, on 2016-January-18, 23:10, said:

Aha, so your system requires playing cuebids strictly as first-round controls. I stopped reading at this point

Would you also have stopped reading about RKCB if this pair of hands were used to promote it?

[1eyedjack]

AyunuS, on 2016-January-18, 22:56, said:

I feel that aces or even keycards do not make slams. They are worth something, yes, but I've never felt it to be all that accurate of a description of the slam capability of a hand.

Blackwood was never intended to identify the slam capability of a hand. I think that you are supposed to ascertain the slam capability first by other means, and then use Blackwood only for the purposes of checking for quick losers.

[mgoetze]

nullve, on 2016-January-19, 01:04, said:

Would you also have stopped reading about RKCB if this pair of hands were used to promote it?

If I were told that I have to use cuebids to show aces exclusively so that I can use RKCB with a void, yes.

[Zelandakh]

AyunuS, on 2016-January-18, 22:56, said:

5♣: 0-3, 10-11, or 18-19
5♦: 4-5, 12-13, or 20-21
5♥: 6-7, 14-15, or 22-23
5♠: 8-9, 16-17, or 24+

1. Your partner has 12-13 such points. You have enough that they can't have 18, and if they did, they probably would've bid higher earlier. And if they've shown even one control they can't be in the 0-3 range.

You apparently do not know your own system yet. The weak option for 5♦ is 4-5, which could be ♦A, ♠K+♦K or ♠Q+♥K+♦K for example. Constructing such hands that would be strong enough for a 3♠ rebid is quite another thing though, perhaps ♠QJT7 ♥KJ9 ♦KQJT98 ♣- would qualify. The problem is working this out at the table - this would (imho) be considerably more difficult than with normal relay methods, so all-told, there would need to be a lot more evidence of the soundness of the method before seriously considering it.

[helene_t]

Where do your 5-4-1-2 coefficients come from? You might want to fit a regression model to find the optimal coefficients. I also feel that the value of a king should mostly depend on whether partner has shortness, could have shortness or does not have shortness. A fit is not required, and also there should be a difference between knowing not to have a fit and not knowing if there is a fit or not.

Maybe you could play that after 3♠, a 3NT bid asks partner to start cuebidding while a 4♣/♦/♥ bid shows slam points according to your scale.

[straube]

Zelandakh, on 2016-January-19, 07:30, said:

You apparently do not know your own system yet.

Why set yourself up like this? His cue bid promises the DA which acts like a key to decoding the range and especially the combinations. All of his example possibilities contain the DA and he must have other cards to be able to jump raise spades. I agree with your point about working out combinations at the table.

[Zelandakh]

straube, on 2016-January-19, 07:46, said:

Why set yourself up like this? His cue bid promises the DA which acts like a key to decoding the range and especially the combinations. All of his example possibilities contain the DA and he must have other cards to be able to jump raise spades. I agree with your point about working out combinations at the table.

The response to that is to take the equivalent hand with the minor suits reversed, presumably with the auction 1♣ - 1♠; 3♠ - 4♦. It is difficult to play a system that relies on a key that is only available some of the time.

[straube]

Zelandakh, on 2016-January-19, 08:13, said:

The response to that is to take the equivalent hand with the minor suits reversed, presumably with the auction 1♣ - 1♠; 3♠ - 4♦. It is difficult to play a system that relies on a key that is only available some of the time.

For sure.

I think this structure lumps things together that ought to be divided. Typically slam auctions proceed...

1) find fit
2) determine combined strength, do we have winners enough for slam?
3) eliminate the possibility of two quick losers in side suits by cue bidding aces as well as kings and shortness. Use a trump cue if available to avoid slam with bad trump
4) eliminate the possibility of a combination of two fast losers (aces) or two unavoidable losers (trump) before committing to slam

As the OP said, we don't really have the distributional information. We also don't know whether we might have a cashing A/K in a side suit before we're up at the 5-level. Then we have to work out a variety of possible combinations of honor cards with a valuation assignment (that seems suspect to me) and then we have to assess the probabilities for slam for each of these.

Appreciate the work involved, but I don't think it will hold up.

[nullve]

AyunuS, on 2016-January-18, 22:56, said:

I'm a person that's never liked blackwood. I feel that aces or even keycards do not make slams. They are worth something, yes, but I've never felt it to be all that accurate of a description of the slam capability of a hand. So, I have created this new little point system of:
A: 5 points
K of a suit with a known fit: 4 points
Any other K: 1 point
Q of a suit with a known fit: 2 points

And then I have created this lovely little system to show how many such points you have:
5♣: 0-3, 10-11, or 18-19
5♦: 4-5, 12-13, or 20-21
5♥: 6-7, 14-15, or 22-23
5♠: 8-9, 16-17, or 24+

Just an observation:

Consider a hypothetical "7A RKCB" that is very much like 6A RKCB except that

* the combination of both non-key (suit) kings is treated as 1 key (suit) queen
* the combination of both key queens is treated 1 one key card
* the Roman-style responses to 4N are, more or less as a result,

5C = 0, 3 or 6 key cards (=> relay => (e.g. step 1 = 0 extra key queens; step 2+ = 1 extra key queen))
5D = 1, 4 or 7 key cards (=> relay => (e.g. step 1 = 0 extra key queens; step 2+ = 1 extra key queen))
5H = 2 or 5 key cards, 0 extra key queens
5S = 2 or 5 key cards, 1 extra key queen

Then, by declaring that

* each key card is worth 4 points
* each key queen is worth 2 points,

the responses can be rewritten as

5C = 0,2,12,14,24 or 26 points (=> relay => (e.g. step 1 = 0, 12 or 24 points; step 2+ = 2, 14 or 26 points))
5D = 4,6,16,18,28 or 30 points (=> relay => (e.g. step 1 = 4, 16 or 28 points; step 2+ = 6, 18 or 30 points))
5H = 8 or 20
5S = 10 or 22 points

or, alternatively, as

5C = 0-2, 12-14 or 24-26 points (=> relay => (e.g. step 1 = 0, 12 or 24 points; step 2+ = 2, 14 or 26 points))
5D = 4-6, 16-18 or 28-30 points (=> relay => (e.g. step 1 = 4, 16 or 28 points; step 2+ = 6, 18 or 30 points))
5H = 8 or 20 points
5S = 10 or 22 points,

where each interval is to be understood as consisting of even integers only.

To make the above scheme look even more like in the OP, we also declare that

* each non-key king is worth 1 point

and change the scheme to

5C = 0-3, 12-15 or 24-27 points (=> relay => (e.g. step 1 = 0-1, 12-13 or 24-25 points; step 2+ = 2-3, 14-15 or 26-27 points))
5D = 4-7, 16-19 or 28-31 points (=> relay => (e.g. step 1 = 4-5, 16-17 or 28-29 points; step 2+ = 6-7, 18-19 or 30-31 points))
5H = 8-9 or 20-21 points
5S = 10-11 or 22-23 points,

where each interval is now a standard integer interval. But despite appearances, this is really the scheme we started with, because each interval

n-(n+1)

just informs us that in addition to the number of key cards and key queens determined by the even integer n, Teller can also have 1 extra non-key king. (Duh!) So here we have a Blackwood-like gadget that looks a lot like the one in the OP, the main difference being just that aces are now worth 1 point less.