"Small" Mistakes Can Cost You A Lot

A few months ago curiosity got the better of me and I investigated video poker. That curiosity was spurned by a losing streak at BlackJack. These happen and when you lose your head, you should step away for a bit. So I did. I'd listened to Bob Dancer over at Gambling with An Edge and his discussion of video poker. I settled on Double Bonus Poker because of the higher payouts for various 4 of a kind hands which come up around 2x per 1000 hands. I learned from Blackjack that going into a game where there is optimal play as opposed to "pull and hope" as is with "reel" slots, without knowledge of how to play would be costly. I had no idea just how costly.

I ruined 10x out of 10. Never experienced anything like that. Even posted to GWAE about it. Asking the dumb, in hindsight question as to whether this was normal.

Ha!

The reason I thought this might be normal is that Double Bonus Poker less than 10/7 pay is a net loss game. BlackJack without a continuous shuffler or other shenanigans is not. So I figured that since my EV was negative and wins and losses vary around EV in the long run I should *expect* to lose many more times than I win. Still, though, I didn't think ruining every. Single. Session. would happen.

When I started playing Blackjack I did not know Basic Strategy but I was learning it. I had an idea that it was a bad idea to hit when the dealer had less than a 7 showing, but I was unawares or all the double down plays outside of 10 and 11, so I gave up quite a bit to the casino on those plays. Still though, with a poor understanding of BS and using Martindale double up strategy I won money more than half the time. Of course I would get wiped out on those losing streaks. But I had a good chance that I would double my buy in. Video Poker was having none of that.

I figured that a game in which you essentially flat bet could not regularly ruin a player. I forgot that if a casino is offering a game it is because it expects to make money off of it. If it doesn't it doesn't put it out. At least not on purpose. The thing with VP is that you can play for houuuuuurs before ruining. Some find that entertaining. Me, not so much. Maybe if I made a long trip to play, but sitting for hours to lose all I brought and to have never been up is not my idea of "fun".

So I did some research and confirmed that I had indeed brought enough so that my risk of ruin was below 10%. So if I was ruining every time, I was definitely doing something very wrong. Well I discovered my mistake. It was one that a beginner or even someone who is thinking but doesn't quite understand the game would do. I won't say *what* I was doing wrong but lets just say that that those plays between low pairs and high pairs are quite significant. The 4 changes I made to my play, which seemed "wrong" but now that I understand the game are obviously right, changed "ruin every time" to 1 ruin in 10. That's a HUUUGE difference in outcome.

So the lesson here for anyone looking for more than entertainment and to not just hand money to casinos is quite clear. Theose "small" mistakes add up and will cost you a whole lot of money. No, the machine is not rigged. Yeah, you may have simply had a bad streak, which are going to happen, but if you are playing a game where there is any strategy or optimal play, you should learn it as much as you can (perfect would be best but we all make mistakes) before putting down the large money. And with the availability of games on IOS and Android, there really isn't a reason to part with your cash until you are proficient.

This Is Not BlackJack

This is a followup to my post about Electronic Table Games. You should probably read that first.

Video "BlackJack"

Have you played one of these:

Have you noticed that something is "not quite right"? Yes. Well I'm about to show you why this Blackjack game is not really Blackjack. Let me explain. In order to understand what I'm about to point out you need to understand a few things:

Blackjack is not simply the rules of the game. Just because you have "decks" of cards doesn't mean you have Blackjack. You have something that looks like blackjack. Blackjack was made for play using physical cards and NOT a random number generator (RNG). Physical cards have certain properties that cannot be changed. They have a thickness and material that sometimes sticks together. I won't go into dealer errors. Not only that but dealing and shuffling of the deck or shoe happens in a certain way.

For example I've watched a dealer shuffle. Here's how it went:

1)The decks are put into a shuffling machine usually referred to as an autoshuffler.
2)A cut card is placed halfway down the deck.
3)The two halves are separated and reversed front to back.
4)The cut card is then placed about 3/4 the way into the deck (also known as penetration).
5)A card is removed from the top of the deck and "burned".
6)The deck is placed in the shoe and the game begins. Note that once the deck is placed in the shoe there is NO OTHER SHUFFLING happening.

Whatever order the cards have, they have until the next shuffle regardless of how many people are playing or when a person enters and leaves the game.

If you've seen my last set of posts you know that there is a particular statistical pattern that will emerge. This is predictable and can be used to "beat" the game so long as you play perfect basic strategy AND use proper bankroll management. Now lets see what it is that these "Blackjack" machines actually do. Please note that I am about to quote the NYS Gambling laws. This differs from other state laws and therefore may not apply where you live. I strongly advise that you find and read the laws governing video games before you put "real" money down.

§ 5319.35. RNG requirements. The selection of game symbols or production of game outcomes for a gaming device using a RNG shall: (a) be statistically independent;
1093 updated (2/17)
(b) conform to the desired random distribution;
(c) pass various recognized statistical tests; and
(d) be unpredictable.

I wish to draw the reader's attention to items a, c and d. Why? Because actual Blackjack is actually predictable and does not "pass various recognized statistical tests". Tell me, have you seen an actual shoe undergo "statistical tests"? No? I thought not. So already we see that video "Blackjack" is not really Blackjack. but it goes further.

An independent testing laboratory approved by the commission may employ the use of various recognized tests to determine whether or not the random values produced by the random number generator pass the desired confidence level of 99 percent. Such independent testing laboratory may choose the appropriate tests on a case-by-case basis depending on the RNG under review. Such tests may include, without limitation:
(a) chi-square test;
(b) equi-distribution (frequency) test;
(c) gap test;
(d) overlaps test;
(e) poker test;
(f) coupon collector’s test;
(g) permutation test;
(h) Kolmogorov-Smirnov test;
(i) adjacencycriteriontests;
(j) order statistic test;
(k) runs tests (patterns of occurrences should not be recurrent);
(l) interplaycorrelationtest;
(m) serial correlation test potency and degree of serial correlation (outcomes should be independent of the previous game);
(n) tests on subsequences; and
(o) Poisson distribution.
[My underlines]

As I have shown in previous posts, in actual BlackJack not only are patterns of occurrences present but are NOT independent of the previous game (where a "game" is any hand). SO lets review:

Video blackjack such as that shown above is specifically programed to eliminate predictability and outcomes that are dependent upon previous hands. That is NOT Blackjack because Blackjack has these two things. But wait, there is more:

§ 5319.37. Background RNG activity requirement. A RNG shall be cycled continuously in the background between games and during game play at a speed that cannot be timed by the player.

It's bad enough that we're using a RNG (which of course has to happen for there to be a "shuffled" deck but unlike an actual deck in which there is no randomization of ANYTHING until an actual physical shuffle occurs, this RNG is constantly in operation. Think about it. Once a Deck hits a shoe the cards are set until shuffle, According to this document, the RNG which actually "picks" the card is constantly running. That is not Blackjack.

The first seed shall be determined randomly by an uncontrolled event. After every game there shall be a random change in the RNG process (new seed, random timer, delay, etc.) in order to ensure that the RNG does not start at the same value every time. It is permissible not to use a random seed. Nevertheless, a manufacturer shall ensure that multiple games will not synchronize.[My underlines]

. Really? I don't know how many people playing a physical shoe have noticed that a shoe was identical but I've never noticed it but lets assume it does happen. Why is a manufacturer required to make sure that something that is possible cannot happen?

§ 5319.39. Live game correlation. Where a gaming device plays a game that is recognizable to be a simulation of a live casino game, such as poker, blackjack, roulette, etc., the same probabilities associated with the live game shall be evident in the simulated game, unless otherwise denoted on the display or help screen.

This is misleading. Number one "same probabilities" doesn't mean what you think it does why? because as quoted above, the games are specifically programmed to avoid predictability and patterns of occurrences influenced by previous games. Again though I have already shown in previous posts that Blackjack by definition is predictable and hands (games) are directly influenced by the hands that preceded them.

Lets look at some other legal requirements:

§ 5319.45. Software requirements for percentage payout.

(b) Gaming devices that may be affected by player skill shall meet the requirement of this section when using a method of play that will provide the greatest return to the player over a period of continuous play.

"...a method of play...". Unless we're talking side bets and IF we're talking side bets....

(c) RNG. Each RNG shall meet the requirements set forth in section 5319.35 of this Part and the following requirements:

(1) In the game selection process:
(i) each possible permutation or combination of game elements that produces winning or losing game outcomes shall be available for random selection at the initiation of each play, unless otherwise denoted by the game;

Again, we know that in a shuffled deck there is no 'randomization" of card picks. That is, the dealer cannot close his or her eyes and pick a card from anywhere in the shoe. No, the shoe was determined at shuffle. There is nothing else to do but pull the cards. It may seem "random" to the player, but the cards are actually set already. You need to understand this. There should be no randomization going on once a deck is created.

Let us also be clear if you have been playing the game pictured above and have been wondering why, for example, you have these long winning and losing streaks. Or if you have noticed that even though certain events happen at a certain frequency almost *all the time*. Or if you have noticed that playing these differs strongly from what you see at a live dealer, you are not imagining things.

Ask yourself the following: The sign says "random shuffle" but have you EVER been notified of a shuffle? No? Why not? If this is supposed to simulate an actual Blackjack game, you should at least be notified that a shuffle has happened. Oh it's when the avatar changes? Really? Did they tell you that? You do know that at a casino a dealer can rotate out between shuffles right?

Say what about penetration. Do those signs ever tell you about penetration? Did you see anything in the NYS Gambling commission about notifying the player about how much of this "random" deck is being used in each game? No? Any real Blackjack player knows that the penetration has an immense affect on the game. You'll note that the laws only require a random deck. It says NOTHING about how much of the deck has to be used. It says nothing in regards to whether the deck has to be the same penetration after every "shuffle" (whenever that is).

[Update] looking over my data for the past few months and comparing it to physical games, I estimate that the penetration on these game ranges from 1/3 to 1/2. Since I have seen no laws regarding notifying the user of penetration (something you can see at a physical game) or of having a set penetration you should assume that the casino can legally change the penetration (or randomize it) in order to increase the house advantage. This statement assumes that a "shuffle" occurs at every avatar change.

[/update] It's not that these companies cannot program such notices. It's not that they cannot show you a cut card and a burn card. It's that it is to the benefit of the casino that they do not.

Now a note about those "continuous shuffle" Interblock games. If you've played against an physical continuous shuffler and researched the mechanism, you know that a continuous shuffler really isn't. It holds cards in something like a 16 card "slot". As each card is used and returned to the deck, it is randomly assigned to a slot. Obviously "full" slots cannot be used. In other words, although the shuffle is continuous, it is not exactly random. For this reason, even physical continuous shufflers can be mathematically mapped for optimum winning play. However these IB games with continuous shuffles have to adhere to the rules laid down by the laws governing gaming. As shown above they do their best to be random and non-recurring and not influenced by previous hands. Again, something that is impossible to do with physical cards and a physical continuous shuffler.

So to close, I want to be clear. I'm not accusing casinos of cheating. Nor am I accusing the manufacturers of cheating. They are offering games under specific rules and laws. They are using computers that are running specific code. You take that risk as a part of "gambling". My only purpose here is to inform the player of what game they are actually playing. IMO these "blackjack" games are specialized slot machines. It is in your best interest to hit and run these machines and stack short term winnings if you are looking to make real money. It would be better to attack a real physical card game where non-randomness and predictability is part and parcel of the game.

Blackjack 3rd “loss” Analysis

So in the previous post I showed the frequency of different win-loss-pushes in a row. I demonstrated that there is a 50% drop in likelihood of a 2nd loss in a row and then another 50% drop in likelihood of a 3rd loss and another 50% drop by the 4th loss. After that the chart levels out with any further loss in a row being equally [un] likely. It also demonstrated that it is a bad idea to keep escalating bets beyond 4 due to this because though it is unlikely that one would lose again, at that end of the scale winning and losing are about equal likelihood (aka: no real advantage).

So the chart shows that the optimal inflection point is between the 3rd and 4th loss. So again used two programs to determine how frequently one hits on the 3rd try (after 2 losses in a row). Same rules applied:

Basic Strategy, no deviations.
6 decks
4 shoes

So since the charts got kinda messed up, the first bar is wins on "3rd try" (just had 2 losses in a row) the second bar is you just lost your 3rd hand in a row and the last bar (3) is you got pushed.

The reason I have 2 charts per app is because I was surprised by the difference between Mobilityware and BJAT. These two apps must be using different algorithms to get their "randomness". I will say that I have found MW to be somewhat streaky. Assuming BJAT to be more accurate you can see that on the second pass it "conformed" to the expectations. It is possible that the first run was an example of negative deviation and indeed I was cash negative for most of that run (was also up 12% at one point and I would have cashed out).

In the charts we can see that the win percentage on BJ Apprentice trainer falls on the low side of wins at 44.7 and 48.27% wins respectively. Admittedly I thought that to be pretty low given the previous charts. I cannot say how this would actually play out but you would “break even” in a give and take situation (assumings your gives aren’t doubles). If you’re fortunate, with a few BJ and doubles on your high bet you can come out far ahead.

Mobilityware showed a far more favorable win rate for the player with 57.14% and 57.8% win rate. This is a SERIOUS advantage and if it played out like that in a casino, one could expect to make a lot of money as that is way beyond the casino advantage.

The average between these two is 51.97% (52%) which is still an advantage over the casino.

Now I have not had practice play that resembled anything I saw at the casino. Some better, some worse, which I guess is about right so take these stats with the appropriate grains of salt and remember these stats are for short term play, I have no idea and have no intentions on how this would work out long term.

BlackJack Win-Loss-Push In A Row Stats

When you are not counting cards you are faced with a situation where you need to use some kind of sound statistical model to win money. Even when you are counting cards you are using a statistical model to approximate your chance of winning. As you may or may not know just because the odds are in your favor of winning a particular hand, doesn’t mean that you will actually win them.

I’ve recently acquired the skill of card counting and proceeded to have the longest string of losses than I ever had. The primary reason for this is that I did not understand that at high counts not only are YOUR odds high of getting “pat hands” and “blackjack” increased but so are the odds of the DEALER getting pat hands and blackjacks AND PUSHES.

So I had to go back to the drawing board. One thing I stopped doing as my earnings increased was “grabbing my winnings” when they were there. That is, I no longer was satisfied with a $50-$100 win for a session ($500 bank). I thought I should and could extend these sessions. It seemed to work that way in practice but at the table, ahhh..no.

Also I looked back at what I did when I did not understand or was able of counting. I ran (and still do) progressions. As indicated in a previous post, running progressions, particularly negative one can deplete your bank very quickly. However there are sound reasonings for why people do run progressions and make money off them in the short term.

Let me emphasize that I am talking short term. I read about players sitting for 20 hours or so. I don’t know how much bank they start with but I have NEVER sat at a BlackJack table for more than 6 hours total. I can barely go an hour before I have to pee. So what follows are relevant win-loss push stats that may be useful for those who cannot count but want to use some sort of statistical model for their betting.

Below you will find two charts. These charts are the results of playing 4 consecutive 6 deck shoes with cut card and dealer hitting on soft 17 while playing basic strategy with NO deviations. Essentially a no counting amateur who managed to memorize or follow BS.

Chart 1 shows the results playing MobilityWare BlackJack 5.5.1 on iPhone IOS 10.2. Chart 2 shows the results playing BlackJack Apprentice Trainer same platform.

What you are seeing are how often “back to back” events occur. Double downs are not considered two hands though one would want to consider the potential monetary impact of a double down loss. Successful splits are counted as double wins. Double loss splits are considered 2 losses in a row and a split split is considered a single loss and single win.

You’ll note that with both apps the odds of winning or losing two hands in a row is around 50%. That is if you’ve lost the first hand the chance of losing the second hand is 50%. You’ll note that winning or losing the third hand halves again the chances. Note that there is a leveling off beginning at 4 hands. There is very little statistical change between win and loss 4 and 8 in a row. This is why if you are playing strictly on win-loss statistics and using a negative progression it is in your interest to stop the bleeding at 4x. At 5x and beyond though the likelihood of losing again is low one has already entered the “damn near impossible” realm and anything goes ( and I have the losses to prove it).

So what does this mean to the basic strategy player with small bank? Well in my opinion, particularly for people playing with small bank (1k and under) you should use the 10-20% rule. When you are up 10-20% of your buy in, bail the table at the next loss and cash out your winnings. Then sit back down and go for another session and repeat. Why do I say this? Well while making these charts I flat betted and was up 100 to 150 on each game and then lost the majority of the winnings. In some cases falling into losing territory (usually on high counts shrug). So there IS money to be made even when flat betting but due to averaging out, the wins that generate that money is going to get wiped out by corresponding losses, you can see that in the chart, there are approximately the same number of wins and losses but losses OUTSTRIP wins in the long term even in this 4 shoe sample. So rather than hand back your winnings to the casino (which you’ll call “leveraging winnings”) cash it out.

That said, knowing these odds, you can (and should) increase your bets per stats. If you’ve lost 4 in a row and continue losing (and you’ve reset your progression right?) you can get to a “large win” if you keep following it. Same rules from counting applies though: Don’t over bet your bank. Keep your cool. One note though. A lot of the damage in negative progressions comes from doubles. You get say 3 losses out and then double on the 4th. If you chase that double by doubling and doubling again as is necessary to level out your bets you run two steps on the chart. I personally say to not chase a double unless it is your first or second loss, consider anything after that a loss to be recovered later if possible.

Also since flat betting can lead to profits, you may want to look at how often you’ve lost or won in a row and throw a larger bet out simply because the stats are in your favor (say, 3rd loss). Don’t blame me for any losses though.

If you count (or have a good eye for tracking when far too many low cards have come out) you can combine these stats and the count to modify your basic strategy play. For instance, on a ridiculously high count if you have a 13 facing a 7, there is a high probability that you will draw a bust card. There is also a high probability that the dealer has a 10 hole card. This means you’re likely to break ahead of the dealer if you draw and still see a 17 OR you’ll stand and the dealer has a 17. But on the chance that the dealer is not sitting on a 17, there is a high probability that the dealer will break on the draw. You might want to let the dealer take the risk rather than yourself.

Another example would be a 15 or 16 against a 7-9 with a high count. Basic strategy says hit the 15 against those and surrender a 16 to a 9. BUT again those high counts mean it’s likely you’ll draw a bust card (usually a 10 value card). Since you’re likely facing a losing situation anyway (stand and lose to a 17=19 or draw for an almost guaranteed bust) why not let the dealer take that risk? But now I’ve strayed into counting territory.

The point here is that Basic (and counting) are long term propositions. A lot of people are not playing for 20 hours straight or even 8 hours straight or going to a casino for an entire weekend. But they want to make money. Money can be made in the short term if you understand what happens short term. If you’re up by 10-20 percent, grab your money! The stats say that if you stay you’ll give it back. And consider this. If you have a disastrous start (something that happened during this test with the BJ apprentice software you have to come back from that before you can profit. So if you’re down 80 and need to get 50 to make your 10% that’s $130 you need to make. That hurts. So look at your down. If you dropped a LOT then you’ve experienced one half of the statistics, the rest will come back (unless you are REALLY unlucky). But when you get back up and you get a little profit, bail and come back. Money management will help you not dig too deep a hole but understand that for a BS player once you’ve gone halfway through your bank (buy in) it’s usually impossible to recoup during that session. It’s probably best to walk away and go home.