ReKO kid: Improved REKO expectation with one index change?

[Parker]

> Doesn't matter much, but I'm curious. Do the KO terms
> "pivot point" and "key count" have
> meaning in reKO? Or would they have the same meaning,
> but just aren't used? In any case, I wouldn't think
> you could call the point where you first increase your
> bet the pivot point.

I believe that the term "key count" was coined by Vancura & Fuchs in Knockout Blackjack, and simply represents the count at which one begins to have an edge. As such, it is an approximation since the actual count at which one has an edge will vary with penetration.

I'm not sure about the origin of "pivot point," although its use predates Knockout Blackjack. All counts, balanced and unbalanced, have a pivot point. The pivot is the point at which running count = true count. (For unbalanced counts this assumes that IRC=(-x)x(#decks) where x is the amount by which the count is unbalanced).

Thus, the pivot point is the point at which we have the most accurate information, since it is unaffected by penetration. This is one reason the KO works as well as it does, since the pivot is equivalent to a Hi-lo TC of +4, at which times we will have a big bet out.

Also note that the pivot point for all balanced counts is 0.

[fatcat519]

> I believe that the term "key count" was
> coined by Vancura & Fuchs in Knockout Blackjack,
> and simply represents the count at which one begins to
> have an edge. As such, it is an approximation since
> the actual count at which one has an edge will vary
> with penetration.

V.&F. said they determined KC by simulation and that it depends almost entirely on number of decks, for any rule set. I assume that Norm would have done the same for his reKO bet schedules, without naming the point at which the bet is raised.

> I'm not sure about the origin of "pivot
> point," although its use predates Knockout
> Blackjack. All counts, balanced and unbalanced, have
> a pivot point. The pivot is the point at which running
> count = true count. (For unbalanced counts this
> assumes that IRC=(-x)x(#decks) where x is the amount
> by which the count is unbalanced).

Not quite clear here. Wouldn't you have to add 4 in the above formula to make the RC=TC?

> Thus, the pivot point is the point at which we have
> the most accurate information, since it is unaffected
> by penetration. This is one reason the KO works as
> well as it does, since the pivot is equivalent to a
> Hi-lo TC of +4, at which times we will have a big bet
> out.

> Also note that the pivot point for all balanced counts
> is 0.

[Parker]

> V.&F. said they determined KC by simulation and that
> it depends almost entirely on number of decks, for any
> rule set. I assume that Norm would have done the same
> for his reKO bet schedules, without naming the point
> at which the bet is raised. Not quite clear here.

However determined, it is still an approximation. If we're halfway through the shoe (or 2-deck pack, or single deck), then it is pretty accurate. If we reach the key count very early in the shoe/pack/deck, it is actually underestimating our edge, and late in the shoe/pack/deck our edge is overstated.

> Wouldn't you have to add 4 in the above formula to
> make the RC=TC?

Only if you want to compare KO TC to Hi-lo TC.

If we apply the formula above to KO, we get IRC=-4 x (#decks). Thus, the IRC will be -4, -8, -24 and -32 for single, double, 6D and 8D respectively. What we are actually doing is adjusting so that the pivot is at 0.

With these IRC's, we can true-count KO simply by dividing by the unseen decks, just as with a balanced count.

Then, as you surmised, a KO TC of 0 would be equivalent, edge-wise, to a Hi-lo TC of +4.

We could even use Hi-lo indices simply by subtracting 4 from them, although it would be more accurate to sim a set specifically for true-counted KO, or TKO as it is sometimes called.