Norm Wattenberger: reKO revisited

[Norm Wattenberger]

> For system for a beginner, splitting 10's adds a lot
> of variance, maybe too much. The other one that
> strikes me as possibly better off deleted is DD 10 vs.
> X, and to a lesser degree, DD 9 v. 7.

I don't think 9v7 is a problem. You have a point with 10v10.

> Also I notice 12 vs. 4 is on the index list. You play
> regular BS at low counts and the index plays after the
> trigger point; you're not supposed to start hitting 12
> vs. 4 at high counts, are you?

My wording error. I said use normal BS at low counts and should have said use one BS. Neither is the normal BS. Actually I may delete that. The problem I have with CBS, for example, is that you are teaching someone an incorrect BS. If they tire of counting; they will fall back on a bad BS. Anyhow, you start standing on 12v4. Not much help in six decks but it does help the DD SCOREs.

[ToAnyOne]

[walkingdood]

Nice work. Are you hitting 12v4 at less than +2?

[Norm Wattenberger]

[Parker]

KO Preferred is the KO count as described in Knockout Blacjack, using the rounded strategy matrix with all strategy deviations occurring at the pivot point, key count, or IRC. Also discussed in the book are KO Rookie, in which no indexes are used, and KO Full, in which more exact indices are used. The authors contend that the gain from KO Full is insignificant.

TKO is True Counted KO, which is not discussed in the book, and is, of course, KO with a true count conversion and true count indices. It is quite powerful for a level one count, out-performing Hi-lo in most games.

[MJ]

> Adding the two Ten splits at +7 would add a healthy
> 5-7%. But, that's not something I like to suggest to
> casual users.

The pivot point with KO is at +4. This is where you have the most reliable information about your advantage. The further you go above or below that value the less reliable your index becomes. I believe the precise index values are as follows:

10s vs 6 split +5
10s vs 5 split +6

Somebody ran the simulations a while back for the 6 deck game.
Are you using a risk averse index or something?

MJ

[Norm Wattenberger]

[Cacarulo]

Before sending my final set I'd been playing with a lot of index variations and the best I came up with for 6D,S17,DOA,DAS,SPA1,SPL3,NS,5/6 is the one you described.
Nonetheless, we can add some new indices to enhance reKO a little more. The problem is that the process is too slow. I've to add indices one by one and see if the SCORE is improved. If it does not then the index is rejected.
The candidates so far are:

13v2
S18vA

A7v2
A6v2
A5v3
A8v4
A4v4
A3v4
A2v4
A2v5
A9v6

We could also try:

A8v3
A9v5

and some splitting pairs (DAS) like:

33v2
44v4
99v7
77v8

I'm pretty sure that with all these indices reKO will outperform Hi-Lo.

Sincerely,
Cac

PS: Will see if I get some time to try some.

[Norm Wattenberger]

Adding indexes doesn't make it that much more difficult when they are all the same value.

[Cacarulo]

> Adding indexes doesn't make it that much more
> difficult when they are all the same value.

That's right and this is what I love of this system. But, we need to be sure that we are not adding the wrong indices.
Also, what is good for 6D doesn't mean to be good for 2D.

Cac

[MJ]

> Adding the two Ten splits at +7 would add a healthy
> 5-7%. But, that's not something I like to suggest to
> casual users.

5-7% increase in SCORE seems to be on the high side. I was thinking along the lines of 1-2% increase in SCORE. Did you run a sim to determine the gain you suggested?

If the gain really is 5-7%, then to exclude the 10 splits is nothing short of gutting REKO. Cacarulo wants to add a ton of index values. KO-P only used a handful of index plays for the 6D game. Basically, REKO is adding a LOT more index plays with only 1 index number as opposed to the 2 index numbers that KO-P used with only a handful of index plays. I'm really not certain if REKO requires less memorization at all! You seem to want less index numbers but more index plays to learn. Seems like a tradeoff to me.

Why include the index plays Cac suggests which probably add virtually nothing to SCORE? I'll say it for the last time: Don't make the same mistake the KO authors made and neglect the 2 powerful 10 splits. Instead, I propose you include them with the RA index and leave out some of the weaker index plays Cac suggests. The net result will be a higher SCORE with FEWER index plays.

If the KO authors included the 10 splits, then KO-P would probably outperform Hi-Lo and more people would be using the system today as a result.

MJ

[Norm Wattenberger]

> 5-7% increase in SCORE seems to be on the high side. I
> was thinking along the lines of 1-2% increase in
> SCORE. Did you run a sim to determine the gain you
> suggested?

Yes I ran five billion rounds with four players.

> If the gain really is 5-7%, then to exclude the 10
> splits is nothing short of gutting REKO. Cacarulo
> wants to add a ton of index values. KO-P only used a
> handful of index plays for the 6D game. Basically,
> REKO is adding a LOT more index plays with only 1
> index number as opposed to the 2 index numbers that
> KO-P used with only a handful of index plays. I'm
> really not certain if REKO requires less memorization
> at all! You seem to want less index numbers but more
> index plays to learn. Seems like a tradeoff to me.

Not true. KO-P has 21 indexes for six decks, Surrender. reKO has 18 and superior performance with Surrender. Cac is adding more to match HiLo. I think that's a good idea as an option.

> Why include the index plays Cac suggests which
> probably add virtually nothing to SCORE? I'll say it
> for the last time: Don't make the same mistake the KO
> authors made and neglect the 2 powerful 10 splits.
> Instead, I propose you include them with the RA index
> and leave out some of the weaker index plays Cac
> suggests. The net result will be a higher SCORE with
> FEWER index plays.

But most counters don't split Tens these days. Whether or not this makes sense is another question and deserves its own thread.

> If the KO authors included the 10 splits, then KO-P
> would probably outperform Hi-Lo and more people would
> be using the system today as a result.

The correct method would be to compare with HiLo Sweet 16 since that's what most use.

[Cacarulo]

> The correct method would be to compare with HiLo Sweet
> 16 since that's what most use.

Actually, KO with exact indices (C22) outperforms Hi-Lo (C22) in 6D. I've posted these sims a long time ago:

6D,S17,DOA,DAS,SPA1,SPL3,NS,5/6,heads-up

       1-4     1-8     1-12    1-16    1-20 

KO     8.62   25.52   34.94   40.69   44.55 

Hi-Lo  9.61   25.83   34.84   40.43   44.24

And this goes without saying that in these sims Hi-Lo is using an exact TC calculation (flooring). Nobody can do this in play but most can play KO without errors. Besides, you ran some sims also showing KO to be better than Hi-Lo with other number of decks.

Sorry, but I'm tired of hearing that Hi-Lo is superior.

Sincerely,
Cac