Treasure in the Snake Pit

In a recent article on this site ("Winning Streaks"), Henry Tamburin gave an excellent analysis of the new and popular streak sidebet. At the end of the article Henry asks:

Why would anyone make a bet with a 0.5% casino edge and then follow it with one that has almost a twenty-fold higher casino edge?

Of course, Henry is right. Under ordinary circumstances no gambler should consider this bet.

Enter the advantage player. The advantage player often has a topsy-turvy view of things. The intelligent gambler assumes, with some justification, that nothing is in the snake pit but snakes. The advantage player knows that the snakes may be protecting something. Whether or not you are prepared to get bitten to find out is a personal choice.

It's possible the streak sidebet may be beaten by card-counting. It's true the win percentage per hand increases a little with the count, though most of the gain is from double-downs, splits, and insurance.

Unfortunately, computer simulations proved this avenue was a dead-end. The bet is not beatable using the strict application of a commercial count system, since the required win percentages to make the bets positive expectation wagers simply do not occur that often.

However, basic strategy or commercial count system departures are clearly sub-optimal when trying to maximize the chance of winning the hand. Also, we don't know offhand if the effects of removal for win percent are that closely correlated with the commercial count systems.

I simmed Hi-Opt I against this game using full strategy departures with no doubling. Obviously, this increases our chance of winning the hand, raising the win percent from 47.5 to 47.8. The targets for the 5- and 4-streak bets are 48.3 and 48.5, respectively.

I discovered the 5-streak bet becomes positive expectation at TC +2, while the 4-streak bet becomes favourable a TC point higher. Increase in win percent is fairly linear, 0.25% per TC point.

So this bet is theoretically beatable, i.e., it offers us some opportunities for making positive expectation wagers, but I did not consider the loss from deviations from basic strategy. Preliminary calculations suggest the compound wager becomes favourable a TC point higher for both 5- and 4-streak bets.

I have not considered the "drag" factor of the extreme volatility of the side-wagers nor the effective loss of penetration in having to wait a number of hands before the bet is resolved.

I am fairly certain that if a complete zero-memory strategy and a count system is developed for this bet specifically, then the bet is beatable as a practical matter. Note that analyzing a side-bet with commercial count systems was a complete blind alley with regard to the very favourable over/under bet.

For more information about blackjack, we recommend: