Purpose:

        Blackjack may be one of the most popular casino games, due in part to its simplicity - score as close to 21 without going over and be higher than the dealer. Supposedly the optimal way to play has been completely worked out and available for all to learn. The result of which is that any deviation from this "perfect" way to play can lead to scorn and scalding from other players. Such deviations may include deciding to stop at 15 or 16 when the dealer's card is a 10 - despite a 55%-60% chance that the next card will lose you the game! The big question is does it even matter anyway? Perhaps the"perfect" way was just put out by the casinos to increase people losing! For this reason, ScientificAmeriKen is putting perfection to the test and will pit this style versus ScientificAmeriKen's own style of making the dealer beat you!

Hypothesis:

        There seem to be a lot of people that stand by the "perfect" way to play -- but as one wouldn't follow a bunch of people jumping off a bridge nor shall ScientificAmeriKen. It is the hypothesis of this experiment that an alternative strategy that forces the dealer to play out every hand will be stronger!

Methodology:

        This experiment will pit two strategies against each other - playing a total of 100 hands for each (50 at a time and alternating between the two stratgies) on 4 different online casinos. The experiments used "play money" provided by the casino. Following the experiment 50 additional hands were played at $0.50 per hand.
        For "perfect play" strategy, the table provided by this online tip site was used (Picture from the website on right). In general, this strategy calls for hitting up to 17 or higher anytime the dealer shows a 7 or higher. Additionally, when you have a pair you can split them, and at most times you can "double down" or increase your bet if you think you can win (for example - when you have an 11 and the dealer is showing a 6 - chances are you will hit 21, while the dealer goes over) - under "perfect" strategy, doubling down is more frequent.
        For the ScientificAmeriKen method, busting (going over 21) is not an option. Therefore, one will hold on a 12 or higher. Splitting happens when the dealer shows a 3-6, and only if the cards are 88 or 99, AA. Doubling down only happens when the player has a 10 or 11, and the dealer shows a 5 or 6.

Results:

        Data on the right represents roughly 400 hands played per gambling style across the 4 websites. In Table 1 we see a breakdown of performance at each gambling website - with win percentages between 40-53% regardless of play style. Next, a breakdown of the percentage of hands where the playing style actually came into play, which was about 26-33%. Of these hands the range of win percentage for "perfect" ranged between 26-45%, while ScienKen style ranged 23-32%. In table 2, all sites are combined revealing a 50% to 46% difference in perfect play versus ScienKen, while when it mattered it was 33% to 30% in favor of "Perfect". The greatest disparity was in total earnings where "Perfect" style won $16,500 while playing $1,000 a hand, while ScienKen style lost $23,000! Finally, as a control to see if "play" money behaves similar to real money, table 3 demonstrates that win percentage and when it matters rates are fairly similar to "play" money.

Conclusions

        Although it is tough for Scientific AmeriKen to justify the equality of the two methods considering the nearly $40,000 difference in winnings, this website is prepared to do just that. The key metrics to be examined are what happened in the first 50 hands always "perfect" versus the second 50 hands ("ScienKen") and subsequent 50 hands (back to "perfect"), the general trend for these is downward decline (click here to view raw data). It is almost as if the online casinos let one feel good in the beginning -- only to slowly suck the life from them! More research would be needed switching the order of when hands are played, or only doing 50 hands a day - or whatever other superstition fits.         A couple other interest aspects from this research. In only roughly 30% of the hands this decision of how to play even comes up, and the total win percentage difference was only ~3 percent -- if this is statistically significant then really, all this bug-a-boo is over a mere 1% difference - The official stance of scientific AmeriKen is thus, "Phewy! let the people play how they want!". Also interesting was the length of the streaks. This is important for players using the Martingale technique, essentially, doubling the bet after each loss to recoup losses. However, starting at a $0.50 cent bet - to withstand the worst streak observed here of 10 in a row, the individual would have to lose $0.50, $1, $2, $4, $8, $16, $32, $64, $128, $256, and then bet $512 to recoup it all back. They would effectively win $0.50 after all that, which is probably not enough to cover the stress induced medical bills.

Note: in order to avoid creating additional gambling addicts like ScientificAmeriKen, the names of the casinos used were withheld. Please use the contact feature if you wish to inquire.