Question on VARIANCE???

I have been noticing that I experience small positive variance and very large negative variance. Is this common? I am not talking about the often referred to observation of negative variation at high counts.

It seems my wins come slowly, and my losses come suddenly. For instance, I doubt I have had this year a single shoe that resulted in a win of five big bets or better. But I can clearly remember at least six shoes with losses of ten big bets or more. For those with a statistical background, my distribution of results has a long fat negative tail, and a short thin positive tail.

Is this normal? I have considered that this could just be a result of the psychological effect of loss aversion, but I am pretty sure that is not it. I have also considered that I could be reversing the sign on negative counts, but have increased my attention to the sign and still see this result. Plus, even if I reversed the sign I should only expect modest additional losses, not substatially increased losses.

Have others experienced this same thing? In otherwords, is this the usual distribution of results, or am I experiencing a fluke, or worse a systematic error?

Is this normal?

No. However, nothing you experience in a short period of play is normal. This is kind of like the fact you are witnessing a miracle every time you see a number plate, because the chances of you witnessing that particular combination are millions to one.

For example, I went through a very long period of play when I appeared to be making losses on ace splits. Is this normal? No. Should I stop splitting aces? Definitely no.

Plus, even if I reversed the sign I should only expect modest additional losses, not substatially increased losses.

Correct. It is just blackjack. Try and get used to it (easier said than done I know...).

If you "reverse the sign" it can be pretty significant. For example, in a DD game I play regularly, my max bet is out at TC=+4. If I botch the sign, and have a max bet out at a -4, that is going to produce some serious negative swings as opposed to cases where the count is + and I bet big...

You ought to play a lot on CVBJ, and at the end of every shoe, or two decks, or whatever game you play, verify that your count is exactly what CVBJ shows at the shuffle point. No point in not at least being right to that point...

If you look at the long-term bell curve expected results, big wins vs big losses should be pretty balanced, remembering that this is a "long term equalization" and that anything can happen short-term.

It is more likely that you remember the busts more clearly than the solid wins...

In several posts not far below, I questioned the value of high counts. Years ago, I won thousands in Vegas spreading $10-30 and $25-75 using RTC. I'm not sure most of my winnings came from high count bettings. But I'm pretty sure my narrow spreads saved me from being wiped out many many times. Had I spread 1-6,1-8 or 1-12 as many pros suggested, I would've been broke at least five times.(not to mention would have been barred everywhere)

Most people's responses to my thoughts are:

1. Anything can happen in a short term. Believe in the long run. Be patient;
2. Selective memories make you only remeber the bad things. Be strong;
3. Thousands softwares have calculated every detail in this game from every angel, there's nothing new can be found;

I feel like repeatedly hearing;
"God has a plan for everything. Short term difficulties are only minor tests (like WWII, 9/11, Tsunami...)"
"If you havn't seen a miracle yet, just blame yourself for not praying hard enough."

I still believe there's more to be found in BJ. I believe casino shoes are VERY different from those any software can simulate. Computer shoes are truly random, while casino shoes are subject to these factors:

1. the dealer collects players hands in a fixed order;
2. 100% of the dealer's moves and over 90% of players' moves are mechanic -- no one hits on hard 17 or higher, no one stays below 12, etc...
3. almost every hand contains at least one card big enough to stop the hand at certain total number;
4. the dealer shuffles in a relatively fixed manner;

All these factors make casino shoes not so random, especially after playing for a while. These casino shoes have their own "rhythm" --- you can hardly find 15 small cards or 15 big cards coming out in a row, while it's not very rare to see from computer shoes. I guess there are as many as 20% computer shoes that are too "random" to ever happen in real casinos.(except on those CSM's) If that's the case, all the study on BJ should be focused only on the 80% possible shoes.

I know the stories of Revere, Uston, M.I.T team....I know millions have been won from BJ tables by counting. When I play in casinos, I still bet more when RTC is +1 or higher, only not spreading more than three times. I'm working on a new approach to find out everything about the "high count mystery". Until then, I will not spread higher than 1-3, neither will I bet big on low counts and minimum on high.

Can't give you a clear answer. That's about all I can think of for now. The important thing is protect yourself. Don't give yourself all out to an idea that's not 100% proven yet, or at least you still have some reasonable doubts about.

That does sound pretty unusual. When you say max bets, do you mean max on each hand? Ie. If your max is 2x$200, and you're saying you had 6 shoes where you got stuck 2 grand, I wouldn't be as concerned as if you got stuck 6 x 4 grand, or 6 x 2 grand with a max of 1 x $200.

One year isn't a very good statistical sample, but I would be concerned if I had a one year loss overall. How many hours have you played, and how do your results compare with your theoretical EV?

ETF

While I understand you were not responding directly to my post, a couple of comments are worthwhile.

1. Counting _is_ a proven methodology to beating the house. It has been verified in theory. By simulation. And in actual practice in a casino. There are other winning approaches such as shuffle-tracking, which is based on counting, or ace prediction which is a different approach altogether.

2. Software +can+ simulate a real house shuffle. Norm's program even lets you specificy the various dealer parameters (CV shuffle) to specify how accurate the riffle is (are the cards perfectly interlaced one by one, two by two, etc, and so forth. But that doesn't mean that a completely random shuffle is any worse or better for the player than a "house shuffle". The house shuffle is "random enough".

3. I've played a significant amount in casinos, and using CVBJ. I don't find CVBJ significantly different from casino play. I don't see long streams of big or small cards frequently in either. "toy" BJ programs I don't know about (the games sold solely for their game value, not for card-counter training).

4. CVCX or CVData would be a good investment if you don't have a good BJ simulator. Either can answer the question "for these specific rules, this penetration, this bet spread, how well (or poorly) will I do? The problem with a 1-3 spread in a 6D game comes in when you play every hand and don't Wong in/out. You play so many hands at no advantage that it takes a big spread on + counts to overcome the .5% loss over the majority of the hands you play. You can beat a shoe flat-betting, but you have to wong in and out and not play every hand...

I'd be happy to run a sim for any set of rules you want and supply the win rate data for a particular bet spread, if you are interested...

regardless of what type of random it is. - You said:

1. the dealer collects players hands in a fixed order;
2. 100% of the dealer's moves and over 90% of players' moves are mechanic -- no one hits on hard 17 or higher, no one stays below 12, etc...
3. almost every hand contains at least one card big enough to stop the hand at certain total number;
4. the dealer shuffles in a relatively fixed manner;

All these factors make casino shoes not so random...

I doubt this is true.

I know what you are trying to say, and I agree that there really is no real difference between a computer derived random deck and a real life shuffled deck. But I take a pretty hardcore view.

Computers do not generate random numbers. They generate psuedo random numbers. This is a well known problem in computer science, and some very very good algorythms have been designed to make computer generated random numbers indistinguishable from true randomness. But they are not random. This is a problem I run into occasionally in my real life work where I model complex financial situations using Monte Carlo analysis. It isn't easy getting a good normal distribution out of excel.

In the extreme I don't even believe anything is random. Randomness is just a way of saying it is too complicated to calculate. Consider a randomly shuffled deck. Now imagine a computer that could record the order of the cards as they enter the discard tray. Then a series of optical cameras that could watch the deck being shuffled. With enough computing power, and a good program, you could calculate the exact order of the cards. So much for a random shuffle.

We have all heard the story of the Croatians, or was it Bulgarians that used a laser and computer to beat Roulette.

And given perfect knowledge of the table and the dice and all wind currents in the room and enough computing power, you can imagine that a robot could be built that could roll sevens every time.

Even weather patterns are not random. Given perfect knowledge and infinite computing power, all weather could be predicted perfectly. Quantum Physics is now claiming that the actions of subatomic particles are not predictable, but only have probabilistic outcomes. This too is a claim I do not believe. It is just a way of saying the problem is too hard to figure out.

Of course, that belief doesn't help me win any more money at blackjack.

This has been a very tough year. I am up just a few thousand for the year in over 250 hours of play. This is well below my expected value, but well within three standard deviations, so it is not an unusual outcome.

Some on this thread have noted it is a small sample size, and it is. I was just asking to see if this was unusual, or if the distribution is skewed in a known way.

Random is defined as being an event in which all outcomes are equally likely. Now as far as we're concerned with dice and roulette and even a blind man playing blackjack. Everything is random is those respects. Yes, you could get a robot to perform the exact same motion and technique to come up with the same outcome, but if you walk into a casino lugging a 2 ton robot behind you, I don't think the casino will let you up to any table LOL. Anyway, we CAN keep track of about where clumps of cards are and be sure to cut so the small ones are moved to the back and just know that you are getting a couple percent better chances.

I've heard the whole pseudo random explanation before, and not that it isn't true, but if nothing is truly random then we need to call up the dictionary people and have them adjust the meaning of random so it fits.

I people say it in real life, movies, TV and such. "I prayed for my brother in Vietnam (or whatever) and he still died" (or something to that effect). "I don't believe in God anymore cause he didn't answer my prayers".

LISTEN UP PEOPLE. HE SAID "NO!!!" He just didn't give you the answer you wanted is what you mean. BOO HOO.

Oh yeah....I need to talk about gambling in my post. Uhhhh.....
Hmmm....

I like Blackjack its cool!!! (wink)

if you can't track every card.

Quantum Physics is now claiming that the actions of subatomic particles are not predictable, but only have probabilistic outcomes. This too is a claim I do not believe. It is just a way of saying the problem is too hard to figure out.

Well, Heisenberg formulated his uncertainty principle in 1927, so this has been the claim going on 80 years now. You might be right, but I'm not sure I want to live in a universe where there's NO possibility of any randomness. Kind of pulls the rug out from under free-will, dunnit?

ETF

To answer your question, no, I don't think your distribution is skewed in a known way. On average, you should have at least as many +10 big-bet-shoes as -10-big-bet-shoes, unless a) you're quitting shoes in the middle with big bets out after hitting a stop-win, b) they shuffle up on you when you win several big bets, or c) they cheat!

There's a very small self-correcting mechanism within each shoe, so wins tend to follow losses and vice versa. But 1) it operates when you're losing as much as when you're winning, and 2) it's much too small an effect to cause what you're experiencing. Luck is big enough. Luck plus a little selective memory is my guess.

ETF

"You ought to play a lot on CVBJ, and at the end of every shoe, or two decks, or whatever game you play, verify that your count is exactly what CVBJ shows at the shuffle point."
Agreed, just as you would if you were dealing cards to yourself, BUT if you find no errors this way you can still not be sure that you are not making errors in the casino.

If you do this right, where you sit down with the TV on, and you play CVBJ heads-up where you can play about 600 rounds an hour, if you can keep up at that speed with no errors, the casino will really seem easy. At least it does to me... I play with CVBJ all the time, so that the 600 rounds an hour feels quite comfortable. When I hit a casino that can do 60-100 rounds an hour, it seems so laid-back and relaxed that keeping up, even with all the distractions, is easy...

Of course nothing will simulate a real casino environment accurately, but if you can get the counting skills, betting skills, and BS departure skills up to that 600 RPH speed, at least you can feel confident that you can handle the business at hand (counting) once you learn to look through all the other stuff...