Hi guys,
I was just checking out MGP's combinatorial blackjack software and ran into a problem that I can't seem to figure out. According to this software, the EV for splitting 10s vs. a 7 is 0.288. According to the WizardOfOdds online table of EVs, it's 0.296. I believe those EVs were generated with Casino Verite. Basically I think these numbers agree. The small difference is probably something to do with the rules or a difference in the technique of estimating split EVs.
The problem is that my own Monte Carlo simulator gives me an EV of 0.509. This is a difference of ~0.2 which is pretty enormous and concerning in my mind.
(BTW -- further investigation shows that my simulator agrees with the other sources when splitting 9s and below but gives completely different EVs for 10s and Aces.)
Since two reputable sources agree about this number, I assume there's a bug in my simulator. But the problem is that my simulator is giving me numbers that seem to agree with common sense, too.
For example, check out the WizardOfOdd tables. All the EVs for splitting 10s are somewhat less than for splitting 9s. How does that make ANY sense? Okay, I thought, maybe it's a question of card composition and taking those 10s out of the deck really skews the numbers. But the WizardOfOdds numbers were calculated using an infinite deck, so that's not it.
The second common sense problem is that if you have a 10 by itself vs. a 7, then the EV of hitting is ~0.26. My EV of 0.509 is basically twice this number, which makes a lot of sense to me. The EV from the other sources is less than 0.3, which makes no sense to me. I mean, you make 0.26 on your first split hand and then somehow only 0.04 on your next split hand? Whaaaaa??
So right now I'm inclined to believe my own numbers but that puts me in the concerning situation of disagreeing with WizardOfOdds, MGP, and I think CV.
If anybody has some insight into what's going wrong here, I'd really appreciate it. Thank you in advance!!
