Systems Archive 1

Re: Suggestion -Count Backwards

Posted by +24 on 23 February 1998, at 2:17 p.m., in response to Suggestion -Count Backwards, posted by MathProf on 22 February 1998, at 12:53 p.m.

I like your method of modifying the true count so that you avoid negative numbers by counting backwards. For beginning counters, I would think this would be almost as confusing as negative numbers. If you start K-O at IRC of 0 you will be counting positive numbers almost all of the time since it is a balanced count which tends to converge towards pivot (in a 6-deck game -- +24). This means the count will be rising as the shoe is depleted.

My initial post was about a method that avoided the need to estimate decks remaining. For novice counters this is no mean feat. Dividing the shoe into quartiles means you are only having to estimate which quarter of the shoe you are in. I think this is much easier to do than traditional deck estimation to the nearest half.

The mean reason for adopting an unbalanced count is to avoid converting to true. If you are going to go to the trouble of developing a system of converting K-O to true (and thus the necessity of estimating decks) why not just stick to a balanced count or switch to Snyder's new "true edge" approach? My approach is a middle ground. For the beginning counter you can start with standard K-O, then progress to dividing the deck into half and having a set of RC numbers that correspond to TC's of +1 to +5 that you memorize for the first half of the shoe and another set of numbers that you switch to for the second half.

After you are comfortable with first half/second half, you can switch to quartiles. I personally think if you are going to divide the shoe into more than quartiles you might as well learn to estimate to the nearest half deck and go all the way for true count conversion.

Olaf confirmed that my approach is valid. My question is: what boost does it give to my EV over the published method for K-O? With a fixed Key Count for the entire shoe as the point at which you increase your bet, you are underbetting in the beginning of the shoe and overbetting towards the end. With my method you are conservatively betting a truncated True Count at running count numbers that have been memorized as the equivalent RC numbers in the quartile for TC's of +1 to +5. This makes your betting progression as accurate as a true count system (maybe more accurate as there is less sloppiness and inaccuracy in quartile estimation than in deck estimation to the nearest 1/2 deck.

I am simply looking to maintain as much of the ease of a running count system with as much accuracy of a true count system as possible. Only sims will tell us how my method stacks up against traditional TC conversion and against no TC conversion as is the case with the published K-O count with a fixed Key Count. Unfortunately, I don't know of a simulator that will allow me to test my method of betting at different running counts depending on which quartile in the shoe you are in. Anybody have any ideas?

Responses