Is your Playing Strategy
Costing you Money?

(This is a rewrite of a proposed article for the BABJ, originally written in Jan. 1994.)

If you read nothing else on Blackjack all year, please, read this!

Do you play multi-deck Blackjack in Las Vegas or Reno? Do you play Blackjack in A/C, Minnesota or on the Chicago-area river boats? Are you still using Basic Strategy or a Thorp/Braun derived card-counting system? (Virtually all card-counting systems are derived from the Thorp and/or Braun computer simulations.) Do you use computer software for your practice play and to determine when you are ready for the casinos?

If any of the above are true, then your strategy is putting your bankroll at risk. Even if you exclusively play single-deck or double-deck Blackjack, you are not immune to the latest casino game control measures. The purpose of this article is give you the low-down on how casinos are subtly but drastically altering the game, and what you can do about it.

Since Sept. 15 1982 (or thereabouts), unbeknownst to most players and indeed most [so-called] experts and professional players, the casino game we know as Blackjack has changed dramatically. These changes effectively increase the house edge anywhere from 1 to 3 percent. You may be thinking, the game has changed? Balderdash! They still use the same rules. In fact, the rules today are more liberal than ever, if nothing else, motivated by competition.

Virtually no strategies generally available today are immune from the latest game control measures implemented by casino management. In fact, it is demonstrably provable that the so-called advanced strategies suffer the most from these counter-measures. To understand what has happened, let's take a brief look at the game and the changes which have occurred.

By now it is well-worn history that Ed Thorp changed the game of Blackjack forever in 1962 with the publication of his book "Beat the Dealer". Casino reaction included such counter-measures as rule changes; preferential shuffling; multi-deck shoes; partial-penetration; restricting the number of hands played, and limiting the betting spread.

Last but not least, casinos implemented the most feared casino counter-measure, the barring of skilled players. Fortunately, this counter-measure is no longer legal in the state of New Jersey. On Sept. 15 1982, New Jersey casinos lost their right to bar skilled-players. Ironically, my one and only barring occurred at the Boardwalk Regency on my birthday, Sept. 14 1982 - I was up a mere $275 at the time. I guess they were getting their last licks in before the deadline.

The Ultimate Counter Measure

Anticipating losing the "right" to bar skilled players, by 1982 NJ casinos began studying legal game control measures that would not only reduce their risk against skilled players, but would ultimately increase their table hold in general, resulting in table profits that are slowly but consistently on the rise.

Game control changes had already begun in 1991 when casinos encouraged the governor to declare a state of emergency allowing them to eliminate the player-favorable early-surrender rule. After the no-barring law became a reality in 1982, a number of experimental game changes were tried, culminating in "Blackjack II", a 12-deck, double shoe game. Because these changes turned out to be cumbersome, essentially all but one were eliminated: casino favorable shuffles.

Shuffle manipulation has all but obliterated Basic Strategy and virtually all card-counting techniques. Surprised? Don't be. But I still win, you may be thinking. My Boris Advanced Count (BAC) still wins too - about 46% of the time on a good day; clearly not enough to beat most of these "new" shuffles.

About 3 years ago I finally admitted that my B-A-C wasn't cutting it anymore, not JUST in the shoe games, but in most of the ShuffleMaster games as well.

The final stage in my awakening came in Sept. 1990. I had spent the summer designing what was to be later known as Version 1 of Boris's Casino Blackjack simulation computer software. Even in 1990 "Boris" was a brilliant piece of software, although I was to soon discover that Version 1 had a fatal flaw.

I designed Boris to be laptop compatible allowing me practice play in my motel room before hitting the casino floor. Unfortunately, during this trip in Sept., I moved from table to table and casino to casino, losing session after session. The Boris Advanced Count was no longer working.

At first I thought my counting was flawed. However, according to Boris my play was right on the money - I was indeed playing "by the book". Thanks to some new relaxation techniques I developed earlier in 1990, I was fairly confident that my losses had nothing to do with playing stress (i.e. "steaming" or "choking-up") during my live play.

I spent several sessions shadow-counting, only to get consistently hammered during excessively high-counts, which are by definition player-favorable. It seemed that every time I had the edge, my bankroll ended up down the drop slot. This was not supposed to be happening.

Out of frustration I aborted this trip and began an inquiry into what was happening. On my way out the door of the Excalibur I grumbled "this is why I quit back in 1987". Less than two weeks later I happened across Jerry Patterson's latest book "Blackjack: A Winner's Handbook". A month later, I stumbled onto his earlier book "Break the Dealer", co-authored with Eddie Olsen. These two books were to change my outlook on the game.

Why Strategies Fail

How could it be that strategies devised by playing millions of hands on the computer could suddenly turn out highly flawed? After all, aren't computers themselves flawless? The computers aren't flawed. The flaws come from one MAJOR assumption extant in Thorp's and Braun's simulations. Because virtually every published strategy is heavily influenced by Thorp and Braun's work (with a minor in Griffin and Epstein), this flaw has mindlessly crept into these strategies as well.

The flaw in Thorp and Braun's work (and virtually all that have succeeded them), is that the cards in these studies were "shuffled" randomly. Never in the history of the game has a casino EVER used a random shuffle. Additionally, in devising multi-deck strategies, theoreticists assumed that multi-deck Blackjack is nothing more than a single-deck game with more decks. This could not be further from the truth.

These seeming minor flaws distorted the outcome of the computer studies. The strategies evolving from these studies are therefore highly suspect. By now (1993) most casino management teams are aware of these inaccuracies. They have translated this awareness into shuffles carefully designed to exploit these flaws.

To understand the impact of these shuffles, let's compare the real world with that predicted by Thorp and Braun. Two pieces of data give us considerable input on the fact that the game has indeed changed.

Real World Situations

Against Random Cards

Avg. Dealer Break

18% - 25%

28.3% - 28.5%

Avg. Dealer Hand

Slightly 19

Somewhat < 19>

Basic Strategy tables and betting recommendations rely on the dealer breaking at least 28.3% of the time. Double/Split recommendations assume a dealer average total of less than 19. When these figures become skewed, the table hold increases dramatically.

Today's casinos more than ever before create a non-random distribution of the cards. Beginning with the way in which the cards are box-ordered (typically A-K,A-K,K-A,K-A) and the way they are "washed" (i.e. prepared before the initial shuffle), cards of like-value frequently clump together: hi-cards with hi-cards and lo-cards with lo. Contemporary shuffle procedures promote a recombination of these hi-card and lo-card clumps.

Play of the cards also serves to encourage like-card or pattern-card clumping. Additionally, many casinos (particularly in New Jersey) have structured the play and card-pickup procedures to augment clumping. During the summer months in Atlantic City, the high humidity and salt air tends to stick cards together, encouraging even MORE clumping.

Has it occurred to you that card-counting RELIES on card-clumping? What do you think causes Hi and Lo card counts: excess Hi and Lo cards in a given area of the shoe. One of the major flaws with card counting is that altho the count may indicate an excess of player-favorable cards, where are they? They could just as easily be located BEHIND the shuffle card, never to be played.

As a case in point, I was watching a game at Caesar's in Las Vegas. The players were getting creamed, so I began to just count the tens. When the shuffle card came out (75%+ penetration) there were still 43 tens left in the remaining unplayed 25% of the shoe. Because Caesar's uses a "trackable" shuffle (they didn't plug the shoe in thiose days), I watched this 25% remain virtually intact after the shuffle. The player given the cut-card unknowingly cut this clump to the back of the shoe. Again the players got creamed. I watch this happen 4 shoes in a row before I finally could not take it anymore and left. What was with these players? Talk about unconscious...

If the cards were TRULY random, running counts would nearly always hover around the zero-mark. There would of course be statistical fluctuations, but these would be the exception rather than the rule. In today's shoe games, on the average only about 1/4 of the cards are anywhere near random. During (approx.) the first 100-150 rounds (it varies) after new cards are brought in, clumping is even more excessive.

I had as much trouble believing all this as you are probably now having, until I programmed a generic "zone" shuffle into "Boris", effectively creating the 2.0 version - I was stunned! My Boris Advanced Count began to lose consistently; the Basic Strategy hands-won-ratio plummeted from approximately 49% to around 42%; the dealer break ratio plummeted to around 23% and the average dealer total was 19+.

In watching the artificially intelligent (A-I) players under Boris, I noticed that these players were table switching more frequently. In Boris Version 3.X, auto-players would table-switch mainly when losing. Examining the statistics for these tables, I could see why: the dealer win-rate was significantly higher than before. Selecting a "random" shuffle made the games more player-favorable again.

By now, you should understand the reason for my losing trips. Even with practice play against Boris before every session (with the unreleased Boris Version 1.1), I was practicing against non-existent conditions. I remedied this situation in May of 1991 by canvassing virtually every strip casino, taking notes on their shuffles and washes, programming these findings into the Boris software.

Because I make frequent casino trips and have numerous "scouts" reporting on shuffle changes, Boris now supports nearly 100 different casinos; and, you can setup your own casino conditions as well. When you select a casino with Boris, you not only get the proper rules and penetration, more importantly you get a near-perfect duplication of their shuffle(s).

Current Boris users tell me that they have saved hundreds and thousands of dollars by not rushing into their favorite casinos until they could first consistently beat that casino with Boris. Before you can consistently win, you must FIRST learn to cut your losses. Casino and table selection is at the heart of this success. Thanks to Boris, I replaced my Advanced count with the Boris Clump-Track procedure (in 1991) and Basic Strategy II (in 1992). I eventually added NBJ, PBJ and Power Blackjack to my playing arsenal. For me, conventional card-counting in the shoe game is pretty much a thing of the past; although, I am open to new ideas.


Indications of Clumping

Earlier, I gave you some statistics describing actual casino conditions. Unfortunately, these percentages and averages are not easy to compute while watching tables in the casino. To avoid losing tables, here are some things to look for: