The assumption that mythematicians like to cling on to and batter us about the head with is that if you are certain to lose more bets than you win in the long run then you must also be doomed to lose more money than you win.
They don't care that models and data summaries covering millions of outcomes with a clear overall house edge include countless thousands of profitable series that consisted of more losses than wins.
They just keep chanting that the sample, however large, was "not representative" and warn that at some point, the sum total of all losses must exceed the sum total of all wins. So, beware, insolent peasants, and be ready to pay your dues.
Here's the latest BST log:
It's true that from time to time, a win-loss pattern will set in that has a succession of isolated wins (wins immediately followed by one or more losses). And if the PB value is outside of your MSL ("do over") range, two wins separated by one loss will not save the day before the cavalry rides over the hill in the form of "twins" (two wins in succession!).
The horrors of long stretches in which the NB keeps jumping skyward with each failed EOS bet diminish dramatically as your confidence grows. But it never hurts to keep in mind that you are playing a game in which the odds are almost always against you.
As I explained in the summary, table limits usually extend the battle to climb out of the hole because without wiggle room in your bet values, the green ceiling forces you to keep betting the max until the WLP switches to a more positive trend.
But that assumes you do not have the option to back away from the layout that has put you in the hole, and resume play at a location where you can continue to strictly follow the target betting rules.
Here are the target betting bets applied by the spreadsheet model to the outcomes in the pencil log shown above. The final totals differ from the numbers at the bottom of the log because the recommended responses to dealer naturals and dealer 21s are not applied. I'll fix that at some point, but my purpose here is to keep on demonstrating that more lost bets does not translate to more lost dough!
The blocks of numbers in the spreadsheet line up with the checks and crosses etc. in the log, so if you print out the three images, it should be fairly easy to match them up.
As always, there were numerous turnarounds that delivered a profit from a negative situation, and overall, the house edge was almost 2.0% gross, confirming that more hands went south than north, so to speak.
The NET AV was +2.79%, demonstrating once again that properly-applied doubles/splits really do deliver a player advantage. We already knew that a 3-2 pay-off is better for the bottom line than a 1-1 pay-off!
Real math (as opposed to mythematics) is reassuringly predictable when it comes to data from games of chance. An analysis of recovered series will invariably show that one third contained fewer wins than losses, one third had an equal number of wins and losses, and one third had fewer losses than wins. Surprise!
The bigger the sample, the more precise the distribution becomes. This set (the outcomes from 090309) made a liar out of me, temporarily, with 45 positive series, 18 negatives, and 21 break-evens, but that's not the norm. The current trial sample (all bets against Ken Smith's BST app) are also out of whack (547/369/315) but the distribution will settle down eventually!
An important reminder: The only person likely to make money out of this blog is you, Dear Reader. There's nothing to buy, ever, and your soul is safe (from me, at least). Test my ideas and use them or don't. It's up to you.
_
No comments:
Post a Comment