Further Thoughts on ShuffleMaster

and the Myth of Random Cards

by Ron Fitch --- a.k.a. Boris

( February 2000)

In previous articles on ShuffleMaster, (1) in addition to explaining how the original ShuffleMaster machines (Sm I, SM II and SM BOX) work, I questioned whether or not these machines really produce random cards; pointing out that the ShuffleMaster's definition of random cards and a Clump-tracker's definition of random cards are quite different.

Some time has passed since the first articled was written, making it ripe to revisit this subject; especially in lieu of the many advances and developments which have occurred in the world of card shuffling and shuffle machines. For starters, let's re-examine the idea of ShuffleMaster.

Casino Benefits of using ShuffleMaster
Overall, casino operations benefit from utilizing shuffle machines at their gaming tables. From the casino's viewpoint (or is it merely machine manufacturer's rhetoric?), the main advantage of ShuffleMaster is the elimination of downtime: time spent by the dealer shuffling cards and not playing hands at the table. Ironically, later in this article we will look at the double-shuffle occasionally used by casinos, which re-introduces down-time.

It is well known by casinos and players alike that down-time penalizes whichever side has the edge in the game being played. Anyone who played Blackjack in Atlantic City during the 1979 "Experiment" (wherein card-counters were allowed to play, but against 50% penetration and limited bet-spreads) know how important down-time becomes. The casinos attributed their 13% reduction in their hold to card-counters winning money; when in fact, by the time the Experiment was about to conclude, most of the card-counters - myself included - had left town; the games were just too tough. (My BIG win for that trip was a mere $220). In actuality, the 13% drop was caused by two things:

  1. the down-time created by more frequent shuffling (they were previously cutting off only 1/3 of the shoe).
  2. The casinos computed their win as a function of the Drop-Hold, when they should have been using the Action-Hold instead. Accuracy of the Drop-hold is easily skewed by players holding onto their high-denomination chips in between playing sessions; as well as, by cashing in after every session - requiring a buy-in to start each new session.

Because the casino is the odds on favorite in all table games (with the exception of Blackjack, and possibly Baccarat), the more hands-per-hour that can be achieved, the greater the casino profits. Using ShuffleMaster promises a hands-per-hour increase of up to 40%. That can result in a significant profit increase; enough to pay for the monthly lease payment on a given machine within the first couple days each month of operation (if it is indeed leased).

Additionally, ShuffleMaster Corp. touts their machines as being a deterrent to card-counters and card-trackers. As we have seen in previous articles and from playing experience, the reverse is often the case. Some games are exploitable not so much because of the shuffle (or lack of it), but simply due to the game rules. To my knowledge, ShuffleMaster does not dictate specific game rules; only the shuffle procedure, as delivered by the shuffle machine itself.

In games like Carribean Stud and Let-It-Ride™ a shoe-like attachment can be added to ShuffleMaster I that "spits" out the cards in groups of 3 or 5 (depending upon the type of game), saving dealers the hassle of counting out the correct number of cards for each player (which is prone to error). Of course, being that ShuffleMaster owns the rights to Let-It-Ride™, they require that a ShuffleMaster unit be "bundled" with every table license; increasing their [ShuffleMaster's] profits. With all this going for the casinos, you would think that there would be a clamor to install ShuffleMaster machines at virtually all of their card table games - yet this is not so.

ShuffleMaster Corp.'s Quarterly report for Q1 of their 1999 fiscal Year (obtained from Edgar Online) is very revealing. Overall (compared to 1998), their Q1 revenues were down 18%, possibly fueled by a 69% drop in (purchased?) Shuffler revenues; which may have been influenced by the 23% increase in unit purchase price. Ironically, ShuffleMaster's service contract revenue increased 43% (required maintenance?) along with a 22% increase in Shuffle Machine leases (after a 14% drop the year before). While the above numbers may indeed be interesting, for purposes of this article, what interests us is machin' use in the game of Blackjack.

You may recall a few years ago the Barbary Coast converted virtually all of their Blackjack tables to ShuffleMaster. The 6-Deck games were supported by the large "ShuffleMaster BOX" (using the "Elevator" method, described in the previous article). For Double-deck games, they replaced their quite lucrative unbalanced shuffle with ShuffleMaster II, presumably to capitalize on the (up to) 40% hands-per-hour increase. Approximately a year later, the ShuffleMaster II's were gone, while the BOXes remained. Apparently, the hands-per-hour increase did not offset the house-win produced by their infamous Double-deck shuffle - Very telling. Additionally, hand-shuffling incurs no maintenance overhead.

Other casino's have eliminated ShuffleMaster in their Blackjack games, while increasing machine use in other table games (such Double-down Stud, Caribbean Stud and 3-Card Poker). Aside from the mechanical problems encountered with these machines, I suspect that the REAL reason ShuffleMaster I & II are being eliminated from handheld Blackjack games is that the cards produced from the machine's shuffle process are HARDLY random, and are in fact quite exploitable. If you have read the previous articles on the subject, you know why this is.

Card-counting and ShuffleMaster
ShuffleMaster claims that one of the benefits of introducing their machines into handheld games (1 and 2 deck) is that the [sic] random cards produced by the machine's shuffle process serves as a deterrent to card-counting. That they make such a claim, in my mind, reveals how little they know about the card-counting process itself. Then again, it ties in with the myth that card-counting requires random cards in order to be effective. This could not be further from the truth. (2)

In single and double deck games, a card counter is looking for a significantly HIGH (+) count in order to place a larger bet; or a significantly LOW (-) count to cut the bet back to whatever minimum is deemed will be effective without drawing heat. In truly random cards, extremes in the count are relatively rare; the count-variance is only a few points plus-or-minus That is why a counter's edge is so small, leaving him/her open to still being wiped-out in the short haul. Amidst non-random cards, the running-count "slope" becomes quite steep; especially in 6 and 8 deck games.

To understand how volatile a counter's alleged-edge REALLY is, I refer you to the Black-bean/White-bean examples in Kenny Uston's books. (3) Casinos share this same volatility, except that they have time and a virtually unlimited bankroll on their side.

You should realize that virtually all card-counting strategies have been devised from computer simulations which utilized random card-shuffling algorithms at the core of the process. When you pit these strategies against ACTUAL casino shuffles, a different picture emerges. Therefore, ShuffleMaster Corp.'s claims that producing random cards serves as a card-counter deterrent tells us that either they haven't a clue about the counting process, or, that their machines really DON'T produce random cards. Which do YOU think it is?

Then again, there is a group of card-counting players (albeit small) who are observant enough to recognize that non-random cards can produce some of the most lucrative card-count opportunities, if you know how to recognize them and play them correctly. In his book Blackjack - A Winner's Handbook, Jerry Patterson details a card-count technique known as the Count-reversal, which is designed to capitalize on the non-randomness of today's games.

Unfortunately, as I pointed out in my review of the book, it is riddled with errors; most notably in the section describing shuffle tracking and the count-reversal techniques. Fortunately, the count-reversal descriptions are accurate enough that a clever mind can extrapolate workable techniques from those pages.

In April 1991, I devised the (as-of-yet unreleased) Boris Clump-count Procedure - otherwise known as the BCCP - influenced by "Winner's Handbook" material. It was reasonably workable, to be abandoned only in favor of today's various Clump-track techniques (altho new research into this method is currently in progress; thanks to Boris). However, encouraged by discussions with a handful of players who successfully combine Card-counting and Clump-tracking in their play, I am again evaluating the BCCP for live-play.

With the BCCP, it should not surprise you to find out that ShuffleMaster II games had my attention in this regard. Unfortunately, the age-old problem of penetration is the major deterrent to Clump-counting in the handheld games; just as it is with Card-counting. If the dealer is giving us barely 50% penetration, that is a TOUGH game to beat. Even IF the cards are somewhat clumped, we run the risk that although the count may rise significantly enough to be exploited, the cards we seek may still never make it into play; giving us only one round of exploitable play.

Even worse, hand-held games are still open to preferential-shuffling. (4) Even if we don't encounter preferential shuffling, consistently betting high in the last rounds of the pack can generate a lot of heat; especially in such "burn-joints" as the aforementioned Barbary Coast. (5)

The World of ShuffleMaster BOX
As I implied in a previous article on the subject, ShuffleMaster BOX creates a different set of challenges. From that article, you will remember that at the beginning of each shuffle procedure (i.e. the insertion of cards into the machine), the BOX "randomly" determines whether to produce a "loose" or "tight" intertwine during the shuffle procedure. Next, it determines how many repetitions (i.e. riffles) to put the cards through (5 or 6). Finally, it generates an un-shuffled "clump" (their terminology, not mine) at the beginning of each riffle-step; essentially duplicating the un-shuffled clump produced by ShuffleMaster I & II.

Studying Boris' simulation of ShuffleMaster BOX, I noticed something that is so obvious it has been overlooked. To understand this, let's review a commonly used shuffle technique; namely, the Straight-Thru shuffle.

Essentially, a Straight-thru shuffle involves breaking the cards down into two stacks, shuffling "picks" from each stack; effectively "merging" them into a third stack. Such a shuffle is often a "Shuffle-tracker's Dream!"; unless of course, it is diluted by Plugging and/or Stripping.

Have you ever watched a dealer repeat a straight-thru shuffle multiple times? If you think back, I'll bet that the dealer repeated the procedure 3 times. The only exception I have seen to this "rule" has been at O'Shea's on the Las Vegas Strip. During Peak Periods they used to utilize only TWO steps. Most likely, O'Shea's reasoning was to cut down on the card-randomization, in addition to getting the cards back into play more quickly during "peak" periods; after all, "no one will notice anyway". (6)

(In the Truth-is-Stranger-than-Fiction section of the website, I tell the story of playing with "The Professor". One of those incidents occurred at O'Sheas, on a Sunday night, against a two-pass Straight-thru shuffle. Currently, O'Sheas consistently uses a 3-pass shuffle procedure and offers some of the worst shoe game rules on the Las Vegas Strip.)

With the O'Shea's exception, when the house uses ONLY a Straight-Thru shuffle, you will most always observe THREE Repetitions of the procedure. If you observe only TWO repetitions, then it is most likely to be accompanied by a Stutter-shuffle. Understanding this is crucial if you wish to tackle the "delicate art" of Shuffle-tracking. Put bluntly, like card-stripping, an odd number of Straight-thru riffles essentially reverses the order of the cards.

Unlike the above-described Straight-thru shuffle, with ShuffleMaster BOX, what is SIGNIFICANT about the number of riffles (is NOT whether they are even or odd, but how MANY riffle-steps are done. Because the BOX uses the SAME intertwine-coarseness for every riffle during the current shuffle "sequence", the number of riffles becomes significant:

  1. Basic Strategy and Simple card-counts seek games wherein the cards are shuffled as much as possible, using a "tight" intertwine.
  2. Clump-trackers, Shuffle-trackers and "Enlightened" Advanced Card-counters seek games with the FEWEST number of riffles and a COARSE intertwine.
  3. An odd number of riffles will find the cards in a relatively reversed order from the discard tray, while an even number of riffles will leave the cards in relatively the same order. Of course which cards are riffled with which cards will have a considerable influence on the cards coming out of the shuffle.
  4. With ShuffleMaster BOX, the card order is "reversed" before the riffle; being reversed back as a function of the card-intertwine process.
  5. Clumps shuffled into clumps produce even more pronounced clumping. Clumps of random cards shuffled into clumps of high cards or clumps of low cards tend to break up those clumps; making them "unreadable".

While we may not be able to know the intertwine selected by the unit, in most cases we CAN count the number of riffles the cards are put through. Clump-tracker's always "qualify" shuffle information based on actual card play, making it easier to deduce the intertwine. Reverse-counters should be able to come to similar conclusions.

Alert Shuffle-trackers, Key-card players and even Clump-trackers can easily exploit the above knowledge. Surprisingly, I have yet to hear of this actually being done. Or possibly, those involved in such activities are wisely being quiet about it.

As a final note, the benefit of such complex tracking can be nullified somewhat if the dealer does not offer the cards for a cut; as has been seen in many casinos sporting the ShuffleMaster BOX. Technically, because the cards are not hand-shuffled, they do not have to be offered for a cut. This is yet another legitimate way to put the cards in play faster. As one dealer explained it to me, cutting the cards can introduce up to a half-minute of additional down-time. When it comes to profit-making, every second counts.

Double-Shuffling with ShuffleMaster BOX
On a play trip to Reno in November 98, I encountered an interesting situation at two different casinos: Dealers were instructed to run the cards through ShuffleMaster BOX twice. The effect of this double-shuffle is that it introduced approximately 2 minutes of downtime; the very thing ShuffleMaster was designed to eliminate. When I asked the dealers why this was being done, I received the same programmed response: "Some guy was caught [sic] cheating last week."

Remember that casino "execs", for the most part, still consider Card-counting and Clump-tracking cheating. If players were INDEED cheating (other than counting or tracking), double-shuffling would do NOTHING to stop it. Despite the double-shuffling, I managed wins in both casinos; although admittedly, I did give-back somewhat.

Being a curious-sort, I ran a series of simulations with Boris, comparing a single ShuffleMaster BOX against the double BOX-shuffle. To provide a more "internal" look at the effects of the double shuffle, I've used the Card-Association screen from Boris' Professional edition. Sample results appear below:

Card Assoc.

Shuffle #1

Shuffle #2

Shuffle #3

Shuffle #4

Shuffle #5

Shuffle #6

Shuffle #7

Sngl Dbl Sngl Dbl Sngl Dbl Sngl Dbl Sngl Dbl Sngl Dbl Sngl Dbl
Low-Low 27.3 27.7 32.0 38.3 30.6 34,8 27.5 31.4 41.2 33.3 23.7 35.9 36.0 36.8
Low-Mid 23.0 26.7 28.5 26.7 31.5 19.9 32.1 30.1 22.5 31.6 28.9 28.4 26.7 28.2
Mid-Mid 38.5 28.0 27.9 33.3 29.6 25.0 27.6 27.0 24.5 27.7 25.6 22.5 26.0 28.2
Mid-Hi 19.0 20.5 21.9 23.9 17.7 21.8 20.0 21.4 25.2 21.8 19.9 25.1 23.9 21.3
Hi-Hi 34.5 40.3 41.2 35.2 45.0 41.0 41.2 39.7 35.9 39.3 33.6 40.8 37.7 39.2
10-10 27.1 30.2 28.1 27.1 35.8 31.4 33.5 31.6 27.1 27.1 25.1 27.5 29.2 30.5

At the very least, the above chart strongly suggests that the extra shuffle significantly alters the random distribution of card-types. If the Box TRULY produces Random cards, there should be little difference between the two sets of data; or at the very least the data sets should not be so uniformly different. It isn't until the 7th shoe that things begin to more rapidly "settle-out", somewhat. Seven shoes is a LONG TIME; easily up to two or more hours of play.

Double Deck and the ShuffleMaster BOX
On another note, in early 1998 I was wandering through the New York, New York casino and noticed an interesting oddity: a ShuffleMaster BOX was being used to shuffle double-deck games. When I asked the Floorman about this, his response was that it saved them money from not having to lease an additional ShuffleMaster II for their double-deck games - an expected response. However, another reason, came to mind; ShuffleMaster BOX might be producing less-exploitable double-deck games than ShuffleMaster II. Setting up a special casino environment with Boris, bears this out. Look for this study to appear in an up-coming article detailing Blackjack Simulation comparison studies performed using the Boris-for-Blackjack software.

Target Factor #22? - The ShuffleMaster Factor
Target players spend much of their time scouting for viable tables. While I have heard Target players claim that they do well in ShuffleMaster games, just as many Target players have an opposite opinion. My gut feeling as to why this is so has to do with not taking into account shoe signatures (detailed in the previous article).

Each of the game-related (as opposed to player-related) Factors need to be qualified as applying to the Red-decks or Blue-decks. Instinct tells me that so-called home-run tables usually exist only for one or the other shoe signatures. Altho home-run tables CAN occur for both shoes simultaneously, this is a rare occurrence; happening well under 2% of the time. Not recognizing this may find you coming in to a "Mirage-table". This happened to me several times, one evening at the old Aladdin. It wasn't until my after-session evaluation of the situation that I realized what I had missed.

ShuffleMaster - Some final thoughts

As you can see from all I have written on the subject, shuffle machines add a number of interesting twists to the game of Blackjack; some being more exploitable than others. Depending upon your point of view, shuffle machines can be a blessing or a curse. Overall, my play against these machines finds me well into the black; although I have been "blasted" a few times (6-deck games at the Gold Coast in Las Vegas and the 8-Deck $10 games at Harrah's-Reno vividly come to mind).

Many a card-counter has lamented the game/shuffle changes introduced in Atlantic City after casinos there lost the right to bar card-counters from play. It is my opinion that they are just crying over "spilled milk". Many players (myself included) love the Atlantic City games; once they adjust their playing skill(s) to MATCH those games.

The principle of adapting your play to the game in front of you applies equally to machine shuffled cards. It is for this reason that the Boris-for-Blackjack Simulation Software includes shuffle simulations for ShuffleMaster I, II and "BOX". If you can win consistently against these Boris' simulations, you stand a greater chance of beating the equivalent games in your favorite casino. Of course, your mileage may vary.

Regardless of which machine(s) are used, penetration is still an important issue. Then again, the problem of penetration can be put to rest forever by some sort of "continuous" shuffling device. As this article is going to print, ShuffleMaster is about to release a pair of shuffle machines (with "Ace" and "King" monikers) to address this issue. Look for write-ups on these new machines in the near future. (See: Shuffle Machines 101).

In the meantime, as Kenny [Uston] would say: "Go Get 'em".

1. See the following articles: The Invasion of ShuffleMaster - Part I and Part II.

2. An interesting comment along this line comes from John Gollehon's book: What the Casinos's Don't want you to Know - P. 117 - last para..

3. See: One Third of a Shoe - (P. 49) and Million Dollar Blackjack - (P. 32)

4. For more information on this see my editorial on preferential shuffling.

5. It has been recently reported that card-counters are still being barred at the BC. In the report I read (Mar. 99), a gentleman was barred even though he had been ahead a maximum of about $50 at one point, and was ahead only by $10 (when asked to leave); with $10 bets, no less!

6. That more shuffling produces more random cards has been known in the card-counting community for close to 25 years; what is surprising is the FEW number of players who have actually, successfully, exploited this knowledge. Jerry Patterson hinted strongly at this point. Quite likely, many readers of the book found themselves bogged down by the book's numerous errors, and moved on to other pursuits.

Back for more Articles          Back to the Home Page