(The original question was posted on bjmath.)
For this game I have the following EV?s
hitting = -.51176
splitting = -.461818 (or yours refined -.4617)
Table 3 yields an almost perfect M (mean) because:
-.51176 + .0499378 = -.461822 a fantastic approach to the true figure.
This is not an approach. By the definition of EOR the mean (m) is calculated as follows:
mean (m) = EV(split) - EV(hit)
so,
EV(split) = mean + EV(hit) which is what you're doing.
A bit of EOR's Theory
If instead of using a 52-card deck you used a 49-card deck (with 8,8 and T removed) then "mean" would be the exact difference between EV(split) and EV(hit). According to the EV tables in bjmath (88vT,1D,S17,DOA,NDAS,SPL1):
EV(split) = -46.1726%
EV(hit) = -51.1755%
mean = 5.0029%
Depending on the cards removed you can have different means. For example:
1) 52-card deck (no cards removed)
mean = 5.0692%
2) 51-card deck (upcard removed)
mean = 5.5999%
3) 49-card deck (T,8,8 removed)
mean = 5.0029%
All these calculations assume that we are using conditional expectations but what if we used total expectations (before the dealer has checked for bj)?
1) 52-card deck
mean = 4.6793%
2) 51-card deck
mean = 5.1607%
3) 49-card deck
EV(split) = -50.5667%
EV(hit) = -55.1612%
mean = 4.5945%
An important thing you should note is that the EORs must sum to zero. This won't happen (against a dealer Ten or Ace) if we use conditional expectations so my advice is to use total expectations.
If you use the same strategy that you used for calculating the mean then your EORs will sum to zero!
Hope this helps.
Sincerely,
Cacarulo
Bookmarks