No, no, no no and no. :-)
Of course it's my fault.
In the shoe you posted Lugi (thanks for your interest) and not taking into account additional important shoe per shoe long term statistical situations, I would have won every hand with my #2 plan.
First BB is followed by BBB (W)
Second BB is followed by B (W)
Third BB is followed by BBBB (W)
This is just a "normal" deviation as a fictional player betting after any B double against another B double is W=3 and L=0
Notice that a second fictional player betting against a BB, BB sequence hadn't the opportunity to bet and the same is true for a fictional player betting against a third BB, BB, BB sequence.
After the cutoff point of BB, BB, BB my plan is over, I'm not chasing or hoping to get situations coming out very unlikely. (Strings of 5, 6 or more B doubles could come out sooner or later.
The important thing is that you consider separately those three fictional players with their W and L situations.
Actually and according to my shoe per shoe findings, I play toward clustered winning situations or after a single losing situation.
To take advantage of the very likely situation to get at least one cluster of P 1s and 2s per any single shoe, you have to wait the first condition to appear. That is an apperance of a P single or P double.
In your shoe, first trigger is single P followed by PPPPPPP (a loss), then the second trigger (another P single) is followed by PP (a win). after that we cauldn't care less of what happens next on the same P1-2 streak.
Do you remember what I've stated about the early P 3+ streaks?
More often than not, they are producing a shoe more rich of such streaks than the average expected ratio (4.5 per shoe). Obviously. It's more likely to get strong deviations after an early strong deviation had come out than the opposite situation (there are intricate card distribution issues that confirm this I do not want to talk about).
Since for our #1 plan P 3+ streaks are very bad, I'm less inclined to put in action this plan even if it would have won after the first L.
Alrelax seemed to agree with that even by considering other aspects.
Notice that with my plans (there are at least a dozen of them) it's far more likely to get a starting W (not here for #1 plan), a WW situation (plan #2), a LW situation (plan #1) as opposed respectively to a starting L, a WL situation and a LL situation.
In addition you see that with my W, WW or LW plan I'm trying to get the best of it not compelling to or forcing the normal expected ratio being W=3 and L=1.
That's because I want to extract a very long winning plan reducing at most the inevitable impact of sequences as WWWWWWWWWWWW or LLLL or LL-LLL-LL that will come out along the way.
Per every starting L, WL or LL events, you expect to get a triple favourable amount of starting W, WW or LW situations. Actually it's even larger than that if the plan dictates to bet banker.
Only the vig reduces the economical return, thus we have to select at most our betting opportunities by a multilayered progression.
I know your interesting point, but I think that a 1:32 plan is much more difficult to manage. Maybe I'm wrong.