I have yet to meet the counter who can play thousands and thousands of hands without any errors. The errors of which I am speaking are strictly playing errors too, not errors in keeping the count (though count-keeping errors would be interesting sim material too). It can be easy to misplay a hand which has a deviation at a large +/- count or to forget to surrender or even to miscalculate the total in your hand and fail to hit or stand accordingly. Or sometimes you just space out right there at the table. I can remember one instance where I zoned out and failed to hit my soft 16 against a T up. And I'm sure I have made errors which I never realized were errors or quickly forgot about. Has anyone attempted to simulate a realistic expectation for someone who makes occasional errors? Perhaps it could be done for one error in every 10, 100, 1,000, 10,000 or even 100,000 hands (the upper and lower values being there strictly for theoretical interest, I'd imagine). Of course, obvious errors like hitting a hard 18, 19 or 20 or standing on totals of 11 or lower or doubling small totals into big cards, failing to split As or 8s, etc. can be eliminated as errors which a skilled player would never make and which would probably make a dealer pause and ask 'Are you sure?'.

It would be interesting to see what kind of effect errors would have on someone's hourly EV and what kind of spread increase would be necessary to overcome the loss associated with various error-rates. If a player can estimate his error-rate, he could probably make use of this type of information.

Also, you hear/read a lot about error-rate being a major factor in the complexity/simplicity debate with regards to how many indices you use or what level of a count you use, but has anyone simulated numbers for these scenarios? For example, comparing the I18 vs. the entire set of indices for a typical count like hi-lo but assuming the player playing all indices has double the error rate?

Anyhow, I'm just curious if any work in this area has been done and what the results look like. Thanks.