Wolverine: Play all spread

[Wolverine]

In my studies of Blackjack, I have come across many thoughts on bet ramping. This site, BJA3, the 2% rule (Kelly), and other publications have all guided my thoughts on betting. While I don?t have simulation software, and I?m sure this can be answered easily there, I am looking for some confirmation of some ideas.

If I lay out the following scenario as the game to be played: play all, a 300 unit trip bankroll to start (it has grown to 350 since the start of 2005) a 40 unit session bankroll, at DD, DAS, s17, 60-65% pen, with one-on-one play.

The 2% rule (Kelly) says anything more than a 6 unit wager is over betting the bankroll, so I have tried to confine myself to a 1-6 spread as follows: 1 unit with any negative count, 2 units at 0 up to +2, 3 at +2 (floored), 4 at +3, 5 at +4, and 6 at +5 TC and above. My first question is: is this considered a 1-6 spread or a 1-3 spread (ignoring the negative counts of 1 unit and starting the spread at 2 units up to 6 units or 3 to 1)?

Because I have read enough to see that spread is a powerful tool, I am tempted to overbet when the count is more than +5 (I recently ran into a DD shoe with a double digit positive TC for the last three deals before the cut card emerged). How much extra risk is there when you get over that 6 unit (7 unit now with a 350 unit bankroll) threshold? Can you minimize that risk by playing two hands at 4 units each (overbetting or good since it is less than the 75% of your max bet (6 units) that BJA3 outlines to minimize risk with 2 hands)?

That is it for now, but I?m sure I?m going to have many more questions as we pursue this line of questioning. Thanks.

[Don Schlesinger]

> While I don?t have simulation
> software, and I?m sure this can be answered
> easily there, I am looking for some
> confirmation of some ideas.

Get the software!!

> If I lay out the following scenario as the
> game to be played: play all, a 300 unit trip
> bankroll to start (it has grown to 350 since
> the start of 2005) a 40 unit session
> bankroll, at DD, DAS, s17, 60-65% pen, with
> one-on-one play.

> The 2% rule (Kelly) says anything more than
> a 6 unit wager is over betting the bankroll,

That isn't true. I'm not sure what you're trying to express here. Your starting unit can be any size; it isn't fixed in stone. If you decide to spread 1-6, your unit for a 300-unit bank will be one size. If you decide to spread 1-8, for the same bank, your unit size will change. So, it simply makes no sense to state, "anything more than a 6-unit wager is overbetting the bankroll."

It's the dollar amount of your max wager that is constrained by your bankroll -- not the number of units it represents. Number of units becomes an issue solely for cover considerations.

> so I have tried to confine myself to a 1-6
> spread as follows: 1 unit with any negative
> count, 2 units at 0 up to +2,

Why 2 units at 0, where you still don't have the edge??

> 3 at +2
> (floored), 4 at +3, 5 at +4, and 6 at +5 TC
> and above. My first question is: is this
> considered a 1-6 spread or a 1-3 spread
> (ignoring the negative counts of 1 unit and
> starting the spread at 2 units up to 6 units
> or 3 to 1)?

It's considered 1-6, but what you call it doesn't really matter; it's whether the bets are close to optimal that matters, and your ramp isn't so great. For the game in question, optimal would be:

TC <= 0: 1 unit
TC = +1: 2 units
TC = +2: 4 units
TC >=+3: 6 units

> Because I have read enough to see that
> spread is a powerful tool, I am tempted to
> overbet when the count is more than +5 (I
> recently ran into a DD shoe with a double
> digit positive TC for the last three deals
> before the cut card emerged). How much extra
> risk is there when you get over that 6 unit
> (7 unit now with a 350 unit bankroll)
> threshold?

If you want to risk a 1-8 spread (cover is the problem), you could bet 8 units (slightly smaller unit size) at +4 and higher, or two hands of 6 units, which increases spread even more, but keeps ROR the same. Instead, why not change the 6-unit bets into two hands of 4 units? A little less risk than one hand of 6, and greater retrurn -- but only if there are other players at the table. Playing alone, as you mention, I'd stick to one hand, because of the card-eating effect of the second hand.

> Can you minimize that risk by
> playing two hands at 4 units each
> (overbetting or good since it is less than
> the 75% of your max bet (6 units) that BJA3
> outlines to minimize risk with 2 hands)?

See above. Not recommended if playing alone.

Finally, 300 units of bank is not optimal for this scenario. You need about 450. So, either up your bank, lower your unit size, or accept ROR of about 23%, which is too high for most folks.

Don

[Wolverine]

I'm not sure I want to do this, but I would like to try to discuss some of these ideas.

> Get the software!
I plan on doing so. But I've seen members confused by the information, and I am very afraid I will be one of them as well. But I see that in order to progress in my learning, I am going to have to. As I read the "output" generated by Zenfighter and others, I am rather confused. I figure I'm in for the same issue with the software.

> That isn't true. I'm not sure what you're
> trying to express here. Your starting unit
> can be any size; it isn't fixed in stone. If
> you decide to spread 1-6, your unit for a
> 300-unit bank will be one size. If you
> decide to spread 1-8, for the same bank,
> your unit size will change. So, it simply
> makes no sense to state, "anything more
> than a 6-unit wager is overbetting the
> bankroll."

I tried to stay away from SPECIFIC numbers of dollars, value of units, etc.. and expressed it as a function of a casino chip unit and table minimums. Since we are going to go down this path, I'll call it a $1,500 bankroll, which divided by a red chip ($5) yields a 300 unit bankroll. I see the $5 wager as the minimum available to bet, hence $5 = one unit. AND, trying to divide an optimal ramp into something smaller than a red chip (unit) makes you look so much like a card counter when you hit those units over and over again. Blending into the tourist woodwork does a card counter wonders! You look systematic enough ranging from 1 to 6 chips only!

> It's the dollar amount of your max wager
> that is constrained by your bankroll -- not
> the number of units it represents. Number of
> units becomes an issue solely for cover
> considerations

Agreed--and the maximum bet should be no more than 2% of the $1,500 bankroll we defined, or $30. Right?

> Why 2 units at 0, where you still don't have
> the edge??

Because of things I have read about the play all approach. Why not start at "2 units" aka, $10 and range down to $5 on all negative counts and range upwards from $10 to your max at $30 as counts get more positive? Hence my later question--is this a 3 to 1 spread ($10 to $30) or a 6 to 1 spread ($5 to $30)?

Secondly, your comp rating is based upon your base wager (2 units or $10) rather than a single unit ($5). This has the potential to double your comps (and since I am not a professional player interested in earning a living, comps become a valuable part of my playing strategy).

> It's considered 1-6, but what you call it
> doesn't really matter; it's whether the bets
> are close to optimal that matters, and your
> ramp isn't so great. For the game in
> question, optimal would be:

> TC TC = +1: 2 units
> TC = +2: 4 units
> TC >=+3: 6 units

> If you want to risk a 1-8 spread (cover is
> the problem), you could bet 8 units
> (slightly smaller unit size) at +4 and
> higher, or two hands of 6 units, which
> increases spread even more, but keeps ROR
> the same. Instead, why not change the 6-unit
> bets into two hands of 4 units? A little
> less risk than one hand of 6, and greater
> retrurn -- but only if there are other
> players at the table. Playing alone, as you
> mention, I'd stick to one hand, because of
> the card-eating effect of the second hand.

> See above. Not recommended if playing alone.

> Finally, 300 units of bank is not optimal
> for this scenario. You need about 450. So,
> either up your bank, lower your unit size,
> or accept ROR of about 23%, which is too
> high for most folks.

That's is the kind of information I needed to hear. Thank you. I will work to put together a bank of 450 units and keep the unit size the same. I will check into the software so I can work on an optimal betting ramp.

At the 450 unit bankroll size you recommend, what is the RoR then?

[Sun Runner]

> Because of things I have read about the play
> all approach. Why not start at "2
> units" aka, $10 and range down to $5 on
> all negative counts and range upwards from
> $10 to your max at $30 as counts get more
> positive? Hence my later question--is this a
> 3 to 1 spread ($10 to $30) or a 6 to 1
> spread ($5 to $30)?

If based on your BR, ROR, and optimal bet ramp your minimum bet is say $10 .. I think coming off the top with that $10 then haphazardly bouncing your bets between $5 and $10 while the
count is < 1 is a great way to play.

If you are playing a shoe game and your max bet is calc'd to be say $150; I would say your defined spread is 1:15 -even though you are making some bets at the $5 level.

What will the pit think? How will they define your spread? I'm not sure but to the casual observer the $5 to $10 bounce should appear erratic.

Maybe this is a good question (as to how the pit sees this) for McDuff.

[Don Schlesinger]

> I'm not sure I want to do this, but I would
> like to try to discuss some of these ideas.

That's what we're here for! :-)

> As I read
> the "output" generated by
> Zenfighter and others, I am rather confused.
> I figure I'm in for the same issue with the
> software.

What is learning all about if not being confused at first and then, slowly but surely, seeing the light? The software is enlightening; you won't be disappointed. I have never heard a single player ever say he was sorry he bought CV or BJRM -- ever.

> I tried to stay away from SPECIFIC numbers
> of dollars, value of units, etc.

No problem. We don't really need them.

> and expressed it as a function of a casino chip
> unit and table minimums. Since we are going
> to go down this path, I'll call it a $1,500
> bankroll, which divided by a red chip ($5)
> yields a 300 unit bankroll.

See, this is one of the problems that we run into over and over again. You get to pick the bankroll or you get to pick the unit size; but, you shouldn't dictate both. Doing so can often lead to being undercapitalized and to a ROR too large to be acceptable. So, in this case, it might make more sense to say, "I want to play $5 units, and I want to spread 1-6. How much bank will I need for x% ROR?" Unfortunately, people don't always think that way.

> I see the $5
> wager as the minimum available to bet, hence
> $5 = one unit.

Fine.

> AND, trying to divide an
> optimal ramp into something smaller than a
> red chip (unit) makes you look so much like
> a card counter when you hit those units over
> and over again. Blending into the tourist
> woodwork does a card counter wonders! You
> look systematic enough ranging from 1 to 6
> chips only!

You can't go smaller than a red-chip unit -- agreed. But, you can, and should, have more than $1,500 of bank for the game you described.

> Agreed--and the maximum bet should be no
> more than 2% of the $1,500 bankroll we
> defined, or $30. Right?

It's not a bad general guideline, but it can vary and be slightly higher, from time to time, with very good rules and/or pen. Again, your bank should be 450-500 units for this game. $2,500 would be much safer than $1,500.

> Why not start at "2
> units" aka, $10 and range down to $5 on
> all negative counts and range upwards from
> $10 to your max at $30 as counts get more
> positive?

I recommend this in the camouflage chapter of BJA3. The house edge at TC = 0 is 0.07%, and you play about 32 hands per hour at that count. Betting $10 instead of $5, means betting an extra $5 x 32, or $160, at -0.07% edge, so doing that will cost you about 11 cents an hour!

On the other hand, there is really no reason to have to do it.

> Hence my later question--is this a
> 3 to 1 spread ($10 to $30) or a 6 to 1
> spread ($5 to $30)?

It's a 1-6 spread. By definition, spread is the ratio of your lowest bet to your highest bet. It doesn't matter where they are placed. If a pit boss is watching, he's going to record that the lowest bet you make is $5 and the highest is $30, no?

> Secondly, your comp rating is based upon
> your base wager (2 units or $10) rather than
> a single unit ($5).

Maybe. It's a good idea to start with $10, but if the count goes south immediately, and you start betting $5 hand after hand, that's what the pit boss might record.

> This has the potential
> to double your comps (and since I am not a
> professional player interested in earning a
> living, comps become a valuable part of my
> playing strategy).

As the saying goes, "couldn't hurt!" :-)

> That's is the kind of information I needed
> to hear. Thank you. I will work to put
> together a bank of 450 units and keep the
> unit size the same. I will check into the
> software so I can work on an optimal betting
> ramp.

Very good idea!

> At the 450 unit bankroll size you recommend,
> what is the RoR then?

If you bet close to optimally, then, by definition, 13.5%. Still too steep? MORE bank!

Don

[Wolverine]

Okay, I get exactly what you are saying about more bank. I really do. And I realize that you can't define all the variables, or there are NO variables except RoR% to fiddle with.

I will also say this: not all of us are pros looking for the ultimate win by counting. I am a serious recreational player. I don't want to lose my shirt, so I count to range my bets when I have an advantage. If I give up some EV to look more like a gambler, then I am willing to do that.

If I come out of a trip having broken even in dollars at the tables, but I got full RFB comps, I think I came out a BIG winner. If I get some cash to build the bankroll more, even better. But, I have to look like a gambler and feel like a gambler to them or they are going to ask me to leave. Play all is my only option given the scenario I usually play under, so I have to plunk out table minimums waiting for the count to go up. If I range my minimum bet from 1 unit up to 2 units occasionally for no other reason than I "feel lucky" then I don't look like I'm just waiting for the positive count to come along. As BJA3 pointed out as well, don't jump your bets just because the count goes up. If you win, parlay. If you don't, hang tight. I realize there are strategies for getting money out on the table (steaming, chasing loses, etc...) when you need to.

Money management and the psychology of BlackJack is the book that needs to be written. The math is great, it gives us the basis for exploiting the game. But if you manage your money wrong, you can count all day long and lose your @ss. Hence my loss limit strategy from investing. Also, the phsychology needs to be fleshed out better. You are *almost* even, and striving for that last 2 units just so you can walk away "even" but then the losing streak starts and instead of cutting your losses and walking away down 10 units, you walk away a 40 or 50 unit loser. From the investment world, making up that kind of loss is a lot harder than the 2 unit loss, no? Why do we all want to get back to even before we leave? It has to be a human psychological phenomenon! Same thing with doubling our money. Raise your hands out there, how many times have you been *this close* to doubling your session bankroll only to walk away "even." Or worse, had to walk away a loser because there was a good count and you lost your max bet 5 times in a row?!?

Have I stepped on any toes out there? I hope so, but in a kind and loving way. The math is fine, but...

...Psychology and money management is where it is I tell ya.

[Don Schlesinger]

> Have I stepped on any toes out there? I hope
> so, but in a kind and loving way. The math
> is fine, but...

> ...Psychology and money management is where
> it is I tell ya.

Suppose you try to double your money but lose back all you've won. Were you going to play tomorrow anyway? Do the cards know what day you're playing?

Either you play with an edge or you don't. And, if you have the edge, and they will let you play, you play.

The rest is all just superstition.

Of course, if you can't take the losses, or get upset by them, then you probably shouldn't be playing in the first place.

Don

[Parker]

You're not stepping on any toes. You're just wrong, that's all.

We play for the long run, so setting any sort of session goals is just silly. You cannot control short-term variance, so why frustrate yourself? In the long run, you will win your total amount wagered times EV.

It's . . . just . . . that . . . simple.

Session results are meaningless, because it's all one session.

You cannot compare blackjack with investing because unlike investing, blackjack can be reduced to pure mathematics.

The only "money management" you need is to be adequately bankrolled for the bet spread you are using. Anything else is superstitious nonsense.

[Wolverine]

See, I stepped on some toes! :-)

(As predicted)

[Garry Baldy]

> Get the software!!

By the way I'm very interested in some software wich can analyze some bizarre betting spreads.

Say 1 box x 1 unit at +1,
then 7 boxes x 2 units at +2,
then 5 boxes ? 6 units at +3,
then 1 box ? 20 units...

...or something like this. It's mostly for camouflage purposes.

Luck.

Garry Baldy.

[bfbagain]

See, I stepped on some toes! :-)

Only Yours, I'm afraid. :-)

(As predicted)

Self-fulfilling prophecy? Seemed like it was a good bet. But hey, we aim to please. :-)

cheers
bfb

[Wolverine]

Thanks for the responses Don and Parker. Let?s get a few things out on the table to see if my ?assumptions? are correct, okay?

I have read somewhere, and I will go find it if I have to, that the player?s edge can be defined as roughly: the amount of + TC at the time of the deal. And, that the average player wins 43% of the hands, loses 49% of the hands, and the other 8% are pushes. If the average hands won is 43%, then obviously the only way to beat the game is to play a strategy which will put more money on the table when you have an advantage (hence bet ramping when counting). Hence, Basic Strategy (BS) evolved to allow players to come near a break even game by splitting and doubling when they have an advantage and minimizing risk when they are at a disadvantage. If there are any discrepancies in the above statements, then please correct them.

Don, you had a comment about playing with an edge. ?If you have the edge, and they will let you play, you play.? That sounds like Wonging to me. And if you can Wong, more power to you. In the play all scenario I am talking about, Wonging isn?t an option. However, knowing when to get up and leave is a very important option the player has. Hence, my money management concerns. Simply stated, if you don?t have the edge (a negative count), then bet 1. If you have the edge, bet 2. The pure mathematics of that indicate that the game has to be totally rigged against you by over a 2-to-1 margin in order for you to not break even, no? If the house wins 67% of the time, and you only win 33% of the time, as long as you bet the 2 at the right time (when you win 33% of the time) you will balance out the 67% (ok, almost balance out) the house wins when you bet 1. Oh, you don?t always WIN when you have the advantage? Why not? Oh, your advantage is very small. So you only expect to win 47% of the time, lose 45% of the time, and push the 8% when the count is positive (I?m guessing at +2 fromt he earlier assumption in paragraph 1, but fill in the correct TC to simulate that advantage) over the LONG HAUL. What happens in any one playing session with that same +2 (?) advantage? Oh, you win 43% of the time, lose 49% of the time, and push 8%. Well within the statistical norms due to the sample size taken (1 hour, 2 hours, 5 hours, X hands? vs the LONG HAUL, aka, infinity). At some point, a human being is going to need to leave the table. You cannot play the long haul ALL IN ONE SESSION! You need to eat, you need to sleep, you need to ?use the facilities? (for real). And when Parker says ?we play for the long run, so setting any sort of session goals is just silly. You cannot control short-term variance, so why frustrate yourself? In the long run, you will win your total amount wagered times EV.? But APs have session goals all the time: play for no more than 1 hour, until you are tired and can?t keep the count, play until you double/triple/quadruple your money, etc? If you set an upper limit or time limit, shouldn?t you set a lower limit (stop loss) too? And Parker just said you can?t control any one session?s variance. I agree, and that is why I?m looking for a method to decide how to play a single session! Parker states, ?Session results are meaningless, because it is all one session.? Sounds good, but is untrue. Please follow along.

Have you ever played 20,000,000 hands in a row at the same table? Is that a big enough sample size to be considered the long haul? Do you need more? Less? Oh, you move from table to table? Have you ever played 20,000,000 straight hands in ?one session?? As soon as you miss a hand (or which you *should* have played, if you are Wonging), that session is over, right? I disagree that a person?s lifetie play is ?all one session.? In statistics, there is a population. We?ll define that as every blackjack hand dealt in the world. You are taking a population sample as soon as you play in LV or AC (because that is where you physically are). Those samples are still large enough that they should mimic the population of all the hands of blackjack played in the world (but I could be wrong). When you go to a specific casino to play, you have further limited the sample down. Even more so, you have chosen a table to play your 20,000,000 hands in a row. You have chosen one table, in one casino, in one city, out of the entire world. And, over those 20,000,000 hands to be played, it is a good bet that the results from that table will correspond to the population sample of all the hands of blackjack in the world due to the large sample size. However, there can also be a statistical variance from that number, no? What happens when you don?t get in those 20,000,000 hands? What if you only get 10,000 in before your session ends? Can the statistical variance be significant enough to wipe out your bankroll? When you are up $1,000,000 after 5,000 hands (of the 10,000), is it time to quit because you have exceeded your EV by such a statistically improbable amount you would be a fool to continue to play (if they will let you, as Don says) Oh, you walk away before you reach $1,000,000 in winnings because you have played for more than an hour? Because the casino is getting edgy about your winnings? Because they will discover you are an AP? Or do you play the remaining 5,000 hands in your 10,000 hand session and lose it all back to even as the statistics normalize for that one table?

Parker says, ?you cannot compare blackjack with investing because unlike investing, blackjack can be reduced to pure mathematics.? I will agree with that statement on one condition, if you can play all the hands in a row. And you can?t. Because each session is an independent sample of the global population of blackjack hands, each session is a small sample. And you stated accurately that the session variance is unpredictable. I equate it very much to the stock market. The overall trend of the stock market is up over time, just as playing AP blackjack has a positive trend over the long haul. However, each blackjack table equates to an individual stock within that market. There are winners and there are losers in both at any given time. And when you play a session at a table in a casino, it is just like buying a stock and then selling it. If you market time correctly in either endeavor, you will do well. But if you market time incorrectly, you will lose, EVEN THOUGH THE OVERALL TREND IS POSITIVE! If you want to broaden the scope a little, instead of buying a volatile stock (which I think is more appropriate for this example), then buy the S&P500 index fund. Again, the trend is up, but if you market time incorrectly, you can break even or lose money easily. Hence, many people believe in the ?buy and hold? technique of investing so that they catch the good and the bad days in the market, trusting that the trend will eventually win over time. But buying a holding for 10 to 20 years is different than playing a ?session? of BJ and then leaving and then starting another session. That does not equate to the long haul. It only increases your sample size. Missing the best 10 to 15 days in the stock market for a year can eliminate any chance of making an investing profit. Missing the best ?times? on a blackjack table should, statistically, do the same thing.

If you manage to leave the table a loser after every session because of poor choices of when you leave (or have to leave), you have effectively timed yourself into losing?even though you play a ?perfect? game of BJ with your betting and index plays! That is why I am asking this question and pushing this issue so hard: what money management and psychological techniques are there to MAXIMIZE your chances of reaching a successful session outcome? If you win every session, even by 1 unit, you will become a winner over the long haul! If you lose every session, by that same 1 unit, you will be a loser over the long haul.

And I?m not even going to get started on my personal observations regarding streaks. You must have played through times when it seems you can?t win a hand, just like you play through sessions when it seems you can?t lose a hand? The sample size is so minimal it is a blip on the radar of the long haul, but in the context of the session you are playing currently, it can make or break you. Haven?t you had streaks like these? Am I the ONLY ONE? ;-)

And we haven?t even hit upon the human shuffle issue! Or the clumping of cards! It may be great to think that BJ is a statistically random game in a given session, where math is the ultimate answer to what is happening, but there is more than just math going on. Hence my curiosity to maximize profits in a given session.

As SSR says, ?I?ll duck and wait for the rotten fruit to be hurled.? :-)

[Parker]

Eventually you will realize that you are a witch doctor in a roomful of MD's, and feel pretty silly about the whole thing.

Or maybe not.

Briefly: If we are talking about 20M hands, it makes not one whit of difference whether these hands were played consecutively at one table or in 20K sessions of 1K hands each. None whatsoever. If you must think in terms of sessions, think of every shuffle as the start of a new session.

You are laboring under the common misconception that if one experiences positive variance, then one is somehow "due" for a negative swing. This is simply not true.

There is nothing new or unique in your post. We have been down this road 1000's of times before. You either accept the math or you don't. The math doesn't care either way. Neither do we.