I realize this is a 'tired, old debate," and I also realize this post probably won't change anybody's mind, but I was reading some excerpts from those who seemed to be experts in statistics. They were discussing inference, and they reminded me that inference is one goal of statistics: to be able to infer information about a population FROM A SMALL SAMPLE. They went on to discuss how the number 30 came to be known as a minimum benchmark for a sample size when using the Bell Curve.
I believe computer simulations are very appropriate when approximating what 'the true nature of reality' will be, given a set of blackjack playing criteria. Yes, I agree that billions of hands should be 'simmed' are long as the simulation criteria match the anticipated, 'real-life' playing session(s) that will follow.
However, when a sim is incapable of accurate prediction due to no fault of the sim itself but rather due to incorrect assumptions made on the part of the programmer, as I believe is sometimes the case with casino blackjack, I believe that the use of the TOOL of statistics, namely inference, is appropriate, albeit inexact, even when employing confidence intervals and critical values, etc.
For those who disagree, what role do you believe inference plays in the field of statistics? Surely we can agree that, for example, when one is trying to predict the winner of a political election, we can agree that polling a sample (not millions or billions) of voters and then infering from this information is acceptable, correct?
I believe computer simulations are very appropriate when approximating what 'the true nature of reality' will be, given a set of blackjack playing criteria. Yes, I agree that billions of hands should be 'simmed' are long as the simulation criteria match the anticipated, 'real-life' playing session(s) that will follow.
However, when a sim is incapable of accurate prediction due to no fault of the sim itself but rather due to incorrect assumptions made on the part of the programmer, as I believe is sometimes the case with casino blackjack, I believe that the use of the TOOL of statistics, namely inference, is appropriate, albeit inexact, even when employing confidence intervals and critical values, etc.
For those who disagree, what role do you believe inference plays in the field of statistics? Surely we can agree that, for example, when one is trying to predict the winner of a political election, we can agree that polling a sample (not millions or billions) of voters and then infering from this information is acceptable, correct?
