Blackjack Tracker
- Thread starter SPX
- Start date
BJinNJ said:
I've thought of using this method to get hold
of 'Beyond Counting'. They won't sell it to you,
but you will get to read it and take notes.
Inter-library loans can take some time, though.
BJinNJ
of 'Beyond Counting'. They won't sell it to you,
but you will get to read it and take notes.
Inter-library loans can take some time, though.
BJinNJ
I’m sure there are still copies around but with such little demand they will be hard to find. We have simulation software now so the book is completely obsolete. It seems like it would be more helpful to have 50,000 randomly generated shoes on paper. That way you could use whatever playing strategy you want and have any number of people at the table. The BJ Tracker book is limited to a full table of people playing BS. You can’t adjust the rules, number of players or playing strategy at all.
-Sonny-
-Sonny-
You might be surprised...
at what different libraries hold.
The NYC library system is huge, as is the Library of Congress.
It wouldn't surprise me if UNLV or the Las Vegas Public Library
collects gambling books. Your local library can request a loan,
then you'll get to borrow it. Your library might have trouble
finding THE library that has your book, though.
That's what takes time.
BJinNJ
at what different libraries hold.
The NYC library system is huge, as is the Library of Congress.
It wouldn't surprise me if UNLV or the Las Vegas Public Library
collects gambling books. Your local library can request a loan,
then you'll get to borrow it. Your library might have trouble
finding THE library that has your book, though.
That's what takes time.
BJinNJ
Try this...
I found the author's address at the bottom of a CasinoCityTimes
article at this link:
http://www.casinocitytimes.com/article.cfm?contentandcontributorid=1354
Maybe you can get a phone number or email address from this info.
Or just write a letter.
BJinNJ
I found the author's address at the bottom of a CasinoCityTimes
article at this link:
http://www.casinocitytimes.com/article.cfm?contentandcontributorid=1354
Maybe you can get a phone number or email address from this info.
Or just write a letter.
BJinNJ
Yeah, but not obsolete if you want to resolve the question or whether or not hand-dealt cards shuffle and flow the same way as simmed cards/shoes. I noticed the other day that in Twenty First Century Blackjack Fred Renzey said he wanted Thomason's results applied to 100,000 hands in order to accept it.
I would think that's a pretty good sampling. Let's take a method like Jay Moore's . . . if I could put it against 100,000 hand-dealt hands from a 6 deck shoe and in the end the system proved to be an obvious loser then I promise I would never again defend the possibility that a progression system might perform differently in the "real world" than in a computer.
It would be enough to satisfy my curiosity and I really do think we OWE it to the blackjack world to settle that dispute and I would be willing to put the hours in to calculate it. The hard thing though is dealing all those hands. I've dealt about a total of 10,000 so far and recorded the results and that has literally taken me months.
With Bob's book I'd be halfway there.
I would think that's a pretty good sampling. Let's take a method like Jay Moore's . . . if I could put it against 100,000 hand-dealt hands from a 6 deck shoe and in the end the system proved to be an obvious loser then I promise I would never again defend the possibility that a progression system might perform differently in the "real world" than in a computer.
It would be enough to satisfy my curiosity and I really do think we OWE it to the blackjack world to settle that dispute and I would be willing to put the hours in to calculate it. The hard thing though is dealing all those hands. I've dealt about a total of 10,000 so far and recorded the results and that has literally taken me months.
With Bob's book I'd be halfway there.
BJinNJ said:
I found the author's address at the bottom of a CasinoCityTimes
article at this link:
http://www.casinocitytimes.com/article.cfm?contentandcontributorid=1354
Maybe you can get a phone number or email address from this info.
Or just write a letter.
BJinNJ
article at this link:
http://www.casinocitytimes.com/article.cfm?contentandcontributorid=1354
Maybe you can get a phone number or email address from this info.
Or just write a letter.
BJinNJ
Thanks a lot for that! I actually ran across an e-mail address for him from a 2000 Casino Times article but was immediately returned as undeliverable. I guess I'll have to try snail mail!
Well I guess I need an education here on what all these people are alleging.
But, if your point SPX, is that hand-shuffled cards are not as "random" as computer sims, I think there's data that supports that. I'm not saying it would make a difference in actual results but I think after hand-shuffling the gap created where the same cards end up is smaller than it would be if completely random. I suppose that's what card-sequencers are relying on to a certain extent? At least I've seen this for single-deck shuffles. Don't ask me where lol.
What are you recording for your 10,000 hands anyway? Whatever it is, I like it lol.
But, if your point SPX, is that hand-shuffled cards are not as "random" as computer sims, I think there's data that supports that. I'm not saying it would make a difference in actual results but I think after hand-shuffling the gap created where the same cards end up is smaller than it would be if completely random. I suppose that's what card-sequencers are relying on to a certain extent? At least I've seen this for single-deck shuffles. Don't ask me where lol.
What are you recording for your 10,000 hands anyway? Whatever it is, I like it lol.
It really comes down to a couple of things. For one, there's the question of whether or not "card clumping" exists in the hand dealt shoe game and can be used for the benefit of the player. A lot of pros either say it's a myth or so the effect is so miniscule as to be useless. There are a few others like Jerry Patterson who push the idea and he devotes a good bit of time to the subject in Blackjack: A Winner's Manual. Say what you will about Jerry Patterson but he's certainly put his time in.
There's also the issue of whether or not the shoe game runs streakier than single deck blackjack. It wouldn't be a stretch of the imagination to believe that this could very possibly be true, since there's the simple fact that there are more of each kind of card in a 6 deck shoe and at least from time to time groups of low cards and groups of 10s will get cozy with each other. (And so I suppose this ties back into card clumping.)
As far as my 10,000 hands . . . I'm probably at about 10,000 right now. I originally dealt out a hundred shoes and that probably took me 2 months or so, working pretty heavily. Now I'm working on my second batch and up to about Shoe # 34.
I hope ultimately to settle this debate once and for all, at least within my own head. I was reading through Walter Thomason's Twenty First Century Blackjack at one point and something struck me. The Foreword is written by Frank Scoblete, who got involved with Walter Thomason somewhere along the line, as did Fred Renzey. In his Foreword, Scoblete says:
"Several gaming writers told him that 5,000 hands does not a study make! . . . Blackjack expert Fred Renzey wanted him to do 100,000 hands. I wanted him to do 20,000. I think experts Henry Tamburin and John Grochowski wanted him to do something in between."
So I chose the highest number: 100,000, which would apparently satisfy all those guys. And if it's good enough for them, it would certainly be good enough for me.
So yes, somehow, some way, at one point I would like to settle the matter once and for all.
There's also the issue of whether or not the shoe game runs streakier than single deck blackjack. It wouldn't be a stretch of the imagination to believe that this could very possibly be true, since there's the simple fact that there are more of each kind of card in a 6 deck shoe and at least from time to time groups of low cards and groups of 10s will get cozy with each other. (And so I suppose this ties back into card clumping.)
As far as my 10,000 hands . . . I'm probably at about 10,000 right now. I originally dealt out a hundred shoes and that probably took me 2 months or so, working pretty heavily. Now I'm working on my second batch and up to about Shoe # 34.
I hope ultimately to settle this debate once and for all, at least within my own head. I was reading through Walter Thomason's Twenty First Century Blackjack at one point and something struck me. The Foreword is written by Frank Scoblete, who got involved with Walter Thomason somewhere along the line, as did Fred Renzey. In his Foreword, Scoblete says:
"Several gaming writers told him that 5,000 hands does not a study make! . . . Blackjack expert Fred Renzey wanted him to do 100,000 hands. I wanted him to do 20,000. I think experts Henry Tamburin and John Grochowski wanted him to do something in between."
So I chose the highest number: 100,000, which would apparently satisfy all those guys. And if it's good enough for them, it would certainly be good enough for me.
So yes, somehow, some way, at one point I would like to settle the matter once and for all.
So exactly what are you recording - wins/losses,ties, BJ's, doubles and splits kind of thing? Number of dealer upcards?
Where a card ends up next shoe compared to prior shoe in a hand-shuffle(which is what I was talking about for a single deck)?
You hand-shuffling 6 decks? If so, you using a consistent shuffle?
By "streakier" you mean number of wins, etc in a row? Lilklihood of winning next hand if dealer lost 2 in a row kind of thing?
Why compare 6 decks to one deck?
Are the results in a spreadsheet?
And, trust me, even if you do a 100,000 it still likely won't be good enough lol.
The more specifically you can state your hypothesis the better off you will be.
Are you using a cut-card? We already know a CSM will deliver wacked out probabilities of high cards compared to a cut-card game, I think anyway lol.
Where a card ends up next shoe compared to prior shoe in a hand-shuffle(which is what I was talking about for a single deck)?
You hand-shuffling 6 decks? If so, you using a consistent shuffle?
By "streakier" you mean number of wins, etc in a row? Lilklihood of winning next hand if dealer lost 2 in a row kind of thing?
Why compare 6 decks to one deck?
Are the results in a spreadsheet?
And, trust me, even if you do a 100,000 it still likely won't be good enough lol.
The more specifically you can state your hypothesis the better off you will be.
Are you using a cut-card? We already know a CSM will deliver wacked out probabilities of high cards compared to a cut-card game, I think anyway lol.
The setup is 6 decks, shuffled according to a shuffle method I found described in the archives here by Fred Renzey (I'd have to find a link). Rules are S17 DOA DAS resplit up to 4 times. And I usually cut out a little over a deck from the bottom.
I deal to 4 "players" and as far as what I'm actually recording, I am recording the actual results of the hands, i.e. whether it was a win or loss, whether it was a split or double down, whether or not the player received a blackjack. The results therefore could be used for a lot of purposes and to draw a lot of conclusions about the characteristics of the game. Particularly for any sort of progression player he/she would be able to see how their progression fared against--presumably, one day down the road--100,000 manually dealt hands.
But it could be used for a lot of purposes. One thing that I'm finding interesting right now is seeing how often a player wins a hand after 3 losses in a row. Consider playing a system in which your method is to flat bet $5s until you hit a 3 loss losing streak and then on the fourth hand you bet $100. If it could be shown that statistically you will win that hand more often than you lose it then that could be considered an advantage.
Some may argue that that's "due for a win" sort of thinking but the truth is that by far most winning and losing streaks occur within a scope of 3 hands and furthermore consider that presumably the reason you're losing hands is because more and more small cards are coming out of the deck and therefore your odds of winning the next hand increase with each loss.
I deal to 4 "players" and as far as what I'm actually recording, I am recording the actual results of the hands, i.e. whether it was a win or loss, whether it was a split or double down, whether or not the player received a blackjack. The results therefore could be used for a lot of purposes and to draw a lot of conclusions about the characteristics of the game. Particularly for any sort of progression player he/she would be able to see how their progression fared against--presumably, one day down the road--100,000 manually dealt hands.
But it could be used for a lot of purposes. One thing that I'm finding interesting right now is seeing how often a player wins a hand after 3 losses in a row. Consider playing a system in which your method is to flat bet $5s until you hit a 3 loss losing streak and then on the fourth hand you bet $100. If it could be shown that statistically you will win that hand more often than you lose it then that could be considered an advantage.
Some may argue that that's "due for a win" sort of thinking but the truth is that by far most winning and losing streaks occur within a scope of 3 hands and furthermore consider that presumably the reason you're losing hands is because more and more small cards are coming out of the deck and therefore your odds of winning the next hand increase with each loss.
OK - I think I'm with you now.
So you're saying 10000 "hands" would be versus 2500 dealer upcards and you're recording the results of each player's hand?
If your results are in a spreadsheet, let me know lol.
It's possible your results may already be so "normal" that playing more wouldn't likely prove much.
As far as progression systems go, you may need a million hands to half convince anybody of anything, more if you vary your bet alot.
Does anybody else think that hand-shuffles may not be "random", compared to a computer anyway, because a card from the prior shuffle isn't, by a long shot, likely to end up the predicted number of cards away from the card it was next to in the first shuffle if they were truly random?
I mean also, hasn't it been shown that if cards are never shuffled, cards will not come out in a random order. Like, because we always split aces, 2 consecutive aces will come out more often in a randomly shuffled shoe than a non-shuffled shoe. So, it's not a stretch to assume shuffled cards may end up somewhere in between.
So you're saying 10000 "hands" would be versus 2500 dealer upcards and you're recording the results of each player's hand?
If your results are in a spreadsheet, let me know lol.
It's possible your results may already be so "normal" that playing more wouldn't likely prove much.
As far as progression systems go, you may need a million hands to half convince anybody of anything, more if you vary your bet alot.
Does anybody else think that hand-shuffles may not be "random", compared to a computer anyway, because a card from the prior shuffle isn't, by a long shot, likely to end up the predicted number of cards away from the card it was next to in the first shuffle if they were truly random?
I mean also, hasn't it been shown that if cards are never shuffled, cards will not come out in a random order. Like, because we always split aces, 2 consecutive aces will come out more often in a randomly shuffled shoe than a non-shuffled shoe. So, it's not a stretch to assume shuffled cards may end up somewhere in between.
So You Want to Study BJ?
<<But it could be used for a lot of purposes. One thing that I'm finding interesting right now is seeing how often a player wins a hand after 3 losses in a row.>>
This has already been done by N. Wattenberg. He just posted a new online
book Blackjack In Color, link below. In this online book he has a chart of the
length of streaks, from streaks of 1 hand to 20 hands.
http://www.blackjackincolor.com/useless3.htm
FWIW
BJinNJ
<<But it could be used for a lot of purposes. One thing that I'm finding interesting right now is seeing how often a player wins a hand after 3 losses in a row.>>
This has already been done by N. Wattenberg. He just posted a new online
book Blackjack In Color, link below. In this online book he has a chart of the
length of streaks, from streaks of 1 hand to 20 hands.
http://www.blackjackincolor.com/useless3.htm
FWIW
BJinNJ
You can save yourself a lot of time by researching this topic before you start dealing all those hands. I believe this topic has been widely covered already:
http://www.blackjackinfo.com/bb/showthread.php?p=60116
And even if you don’t want to rely on those results you should at least research the methods for analyzing shuffles and calculating randomness before you start your experiment. What happens if you deal 100,000 hands and find out that you didn’t record all the information you need for your analysis? What if you realize that you only needed to shuffle the cards, not actually deal them out? What if you do all that work but don’t know what to do with the data? You’ve just wasted a lot of time for nothing.
You need to start with a solid plan first. Research the techniques that others have used first, then decide what methods you will use for your own analysis. Don’t start playing with those cards until you know exactly what you will be studying and how you are going to calculate your results. It will save you a lot of time and effort.
-Sonny-
http://www.blackjackinfo.com/bb/showthread.php?p=60116
And even if you don’t want to rely on those results you should at least research the methods for analyzing shuffles and calculating randomness before you start your experiment. What happens if you deal 100,000 hands and find out that you didn’t record all the information you need for your analysis? What if you realize that you only needed to shuffle the cards, not actually deal them out? What if you do all that work but don’t know what to do with the data? You’ve just wasted a lot of time for nothing.
You need to start with a solid plan first. Research the techniques that others have used first, then decide what methods you will use for your own analysis. Don’t start playing with those cards until you know exactly what you will be studying and how you are going to calculate your results. It will save you a lot of time and effort.
-Sonny-
Of course clumping exists. That’s how card counters get an advantage. They wait for clumps of low cards to come out and bet high into clumps of high cards. Unfortunately, progression systems do not track the clumps, only the win/loss ratio of a single player. That information does not reliably indicate any change in advantage or the composition of the shoe.
But that’s just your hand. What if everybody else at the table gets a blackjack? Just because you lost the hand doesn’t mean that overall more low cards have come out. Progression players are always trying to find ways of tracking the cards without having to track the cards. How do they expect to achieve something that they are specifically trying to avoid?
Sure, he talks about it a lot but he never gives any evidence to support it. That should be the first sign of trouble. His system is based on a presumption that he never bothers to validate. A little research will turn up numerous articles on Arnold Snyder’s website that explain exactly why the TARGET system is ineffective:
http://www.blackjackforumonline.com/content/TOClibrary.html
As BJinNJ pointed out, this has been answered already on Norm’s website.
The experiment you are doing now is only accurate for that one situation. Again, I think it would be much faster and much more worthwhile to just record the composition of the shoe after each shuffle. Once you have the order of the cards set up you can program a computer (or play it on paper) to play the hands however you want with any number of players using any strategy (almost nobody plays proper basic strategy all the time, not even card counters) with any set of casino rules. Also, it would allow you to analyze the frequency of each card and see if each card truly is random or not (and to what degree). You could run a Chi Squared test to analyze the distributions and maybe run a few of the Marsaglia Diehard tests to check for randomness. That information would be much more valuable to more people.
-Sonny-
But that’s just your hand. What if everybody else at the table gets a blackjack? Just because you lost the hand doesn’t mean that overall more low cards have come out. Progression players are always trying to find ways of tracking the cards without having to track the cards. How do they expect to achieve something that they are specifically trying to avoid?
Sure, he talks about it a lot but he never gives any evidence to support it. That should be the first sign of trouble. His system is based on a presumption that he never bothers to validate. A little research will turn up numerous articles on Arnold Snyder’s website that explain exactly why the TARGET system is ineffective:
http://www.blackjackforumonline.com/content/TOClibrary.html
As BJinNJ pointed out, this has been answered already on Norm’s website.
The experiment you are doing now is only accurate for that one situation. Again, I think it would be much faster and much more worthwhile to just record the composition of the shoe after each shuffle. Once you have the order of the cards set up you can program a computer (or play it on paper) to play the hands however you want with any number of players using any strategy (almost nobody plays proper basic strategy all the time, not even card counters) with any set of casino rules. Also, it would allow you to analyze the frequency of each card and see if each card truly is random or not (and to what degree). You could run a Chi Squared test to analyze the distributions and maybe run a few of the Marsaglia Diehard tests to check for randomness. That information would be much more valuable to more people.
-Sonny-
Last edited: