Yes, quite understandable. With a dealer 4 at a very high count, the jokers in the shoe will be the 7s which aren't tracked in Hi-lo. On balance of probability, there'll be a sufficient number waiting to pop out to marry up with a probable 10 - leaving the house to clean up on a double max unit bet. I think I shall cast this one into the abyss as recommended.
On a simiar line of thought, would it be prudent to deviate from BS and take a more conservative approach to doubling down against 2,3,4s at higher counts (+5, +6, +7, +8)? Doubling down say a 9v3 is going to carry even a higher risk, bearing in mind the possible conbinations of non counted cards and 10s that could pop out? I think one of those 5 million hand sims needs to be run, where doubling against dealers 2,3,4s is replaced with hitting, to see what the alteration in the EV would be. Anyone up for it? - I don't have the software.
I'm assuming that BS was produced using unadjusted (for the count) mathematical probabilities based on a flat-betting approach. Where play is based on spreading 1-8/16 (with the max bet out at +4) this must have an effect on the level of risk where the BS doubling plays against 2,3,4s are made at high counts? ie the potential cost of losing a double max bet (16/32 units) is greater than the gain of winning the hand on a single unit basis? Am I explaining the dilemma OK?
Newb99