Perhaps a silly question but...

[Fun_at_21]

...how would long run advantages/disadvantages be calculated if someone were counting SOME of the time, but simply playing basic strategy most of the time?

The one thing I CAN gather is that the long run edge could never be WORSE than the .40 - .75 of a percent disadvantage that a constant basic strategy player plays with to begin with. Obviously, ANY counting can only improve that disadvantage (over the long run).

My guess is this is a very bet-size dependant thing. Naturally, things seem to always be discussed as "those who count and those who dont". But I rarely hear or read of what long term edge might, or might not arise from those who might only (correctly) count now and then, while otherwise just playing it safe and utilizing proper basic strategy? Perhaps these players might be considered sincere but "free lance" counters? They dont count long enough to draw suspisions or to tire their own minds but just to ocassionally draw the house edge closer to even every now and then.

My guess is this very bet-size dependant? But, in theory, couldn't even sporadic/infrequent counting still lower the house edge, especially if these random bets were bigger while the normal basic strategy bets were small?

I also realize that the METHOD of sporadic counting that is used would surely make a difference as well. Stronger methods might require less (or smaller bets) to make more of an occasional impact.

At any rate, I don't know if I'm even making sense here. I guess I was just mostly curious how a long run advantage/disadvantage might be calculated for those who count in isolated fashion rather than as a method...

 

[SleightOfHand]

...how would long run advantages/disadvantages be calculated if someone were counting SOME of the time, but simply playing basic strategy most of the time?

The one thing I CAN gather is that the long run edge could never be WORSE than the .40 - .75 of a percent disadvantage that a constant basic strategy player plays with to begin with. Obviously, ANY counting can only improve that disadvantage (over the long run).

My guess is this is a very bet-size dependant thing. Naturally, things seem to always be discussed as "those who count and those who dont". But I rarely hear or read of what long term edge might, or might not arise from those who might only (correctly) count now and then, while otherwise just playing it safe and utilizing proper basic strategy? Perhaps these players might be considered sincere but "free lance" counters? They dont count long enough to draw suspisions or to tire their own minds but just to ocassionally draw the house edge closer to even every now and then.

My guess is this very bet-size dependant? But, in theory, couldn't even sporadic/infrequent counting still lower the house edge, especially if these random bets were bigger while the normal basic strategy bets were small?

I also realize that the METHOD of sporadic counting that is used would surely make a difference as well. Stronger methods might require less (or smaller bets) to make more of an occasional impact.

At any rate, I don't know if I'm even making sense here. I guess I was just mostly curious how a long run advantage/disadvantage might be calculated for those who count in isolated fashion rather than as a method...

You are right on the most part. Assuming you are counting correctly and playing perfect BS when not counting, your expected win/loss will never be less than your expected loss playing BS or greater than your win from CC. Combining the 2 will generate an overall edge depending (as you said) on your action on the two games.

As far as long term (dis)advantage, you can just find your average bet in the CC game and the BS game, as well as the amount of hours you played (hands would be even better). Then you can do overall edge = ( CC avg bet * CC hours * CC advantage + BS avg bet * BS hours * BS advantage ) / ( CC avg bet * CC hours + BS avg bet * BS hours )

 

[Fun_at_21]

You are right on the most part. Assuming you are counting correctly and playing perfect BS when not counting, your expected win/loss will never be less than your expected loss playing BS or greater than your win from CC. Combining the 2 will generate an overall edge depending (as you said) on your action on the two games.

As far as long term (dis)advantage, you can just find your average bet in the CC game and the BS game, as well as the amount of hours you played (hands would be even better). Then you can do overall edge = ( CC avg bet * CC hours * CC advantage + BS avg bet * BS hours * BS advantage ) / ( CC avg bet * CC hours + BS avg bet * BS hours )

Thanks for the formula.

I attempted to plug in some numbers, just for fun, but based it off a one-time visit playing mostly basic strategy and a small amount of counting. I assumed it was just me and the dealer and the one time visit only consisted of about a half hour of play time. More specifically, I did estimate HANDS played for this visit rather than hours.

I used "conservative" bet amounts for the example and even for the counting method (I actually assumed the counting method to only provide a -.10 disadvantage).

The factors I assumed were...

Avg CC bet = $12
Hands played in CC = 15
CC advantage = -.10
Avg BS bet = $6
BS hands played = 100
BS advantage = -.45

So, plugging these into the provided formula I ended up with...

12 x 15 x .10 +6 x 100 x .45 = 288 / 12 x 15 + 6 x 100 = 780

288 / 780 = 0.36923

If I did that correctly it looks like my above one time visit example would result in an expected combined (dis)advantage of about -.36 of a percent. Or about .09 better than if I hadn't done ANY modest counting and just stuck with pure basic strategy. Given my very small and conservative example, that seems about right to me I suppose.

Assuming I figured everything properly?

 

[assume_R]

Avg CC bet = $12
Hands played in CC = 15
CC advantage = -.10
Avg BS bet = $6
BS hands played = 100
BS advantage = -.45

So, plugging these into the provided formula I ended up with...

12 x 15 x .10 +6 x 100 x .45 = 288 / 12 x 15 + 6 x 100 = 780

288 / 780 = 0.36923

If I did that correctly it looks like my above one time visit example would result in an expected combined (dis)advantage of about -.36 of a percent.

I know I'm nitpicking here, but while you did the calculation correct, you wrote the formula wrong in your text . Remember that -.10 is actually -.10%, or -0.001. And it's a disadvantage, so it should be negative. This matters when your CCadvantage is positive and BSadvantage is negative. And don't write (something = 288) / something else. Makes it seem like one big formula, and is mathematically incorrect. Again, sorry for nitpicking but I just want to make sure you're clear on the whole process.

So you calculated it correctly, but it should have been written as:

($12/hand x 15hands x -0.001 + $6/hand x 100hands x -0.0045) / ($12/hand x 15hands + $6/hand x 100hands) = (-$2.88) / ($780) = -0.00369 = -.369%

If you estimated your CCadvantage at +1.5%, and played the same hands at that advantage, it would be:


($12/hand x 15hands x 0.015 + $6/hand x 100hands x -0.0045) / ($12/hand x 15hands + $6/hand x 100hands) = ($0) / ($780) = 0% advantage (break even).

Nice job on the calculation though!

 

[assume_R]

...how would long run advantages/disadvantages be calculated if someone were counting SOME of the time, but simply playing basic strategy most of the time?

Also, realize that most of your advantage doesn't come from not playing basic strategy, but rather by resizing your bets (except for some situations in pitch games... but disregard that). So in both situations you could be considered "counting", and also playing "basic strategy." Only when you use indices are you considered to not be playing basic strategy.

In essence, you probably meant "if someone were spreading some of the time correlated with the count, but simply flat betting most of the time."

Whether or not you are deviating from basic strategy simply adds a little bit to your EV, whether or not you are spreading your bets. If you are playing for entertainment and comps, you can reduce the house edge by using indices (i.e. deviating from basic strategy), but not spreading your bet and hence still be at a disadvantage, albeit slightly less disadvantage. Make sense?

 

[Fun_at_21]

Also, realize that most of your advantage doesn't come from not playing basic strategy, but rather by resizing your bets (except for some situations in pitch games... but disregard that). So in both situations you could be considered "counting", and also playing "basic strategy." Only when you use indices are you considered to not be playing basic strategy.

In essence, you probably meant "if someone were spreading some of the time correlated with the count, but simply flat betting most of the time."

Whether or not you are deviating from basic strategy simply adds a little bit to your EV, whether or not you are spreading your bets. If you are playing for entertainment and comps, you can reduce the house edge by using indices (i.e. deviating from basic strategy), but not spreading your bet and hence still be at a disadvantage, albeit slightly less disadvantage. Make sense?

Thanks assume_R, that does help to better clarify what I was trying to say. I figured the bet sizing was the key to how much benefit any minimal counting might be.

And thanks for the help in clearing up the formula layout. Indeed, I first wrote the formula out with the negative in front of the percentage but thought it looked "odd" because I had a multiplication sign running into a minus sign...LOL. So I ended up leaving them off and adding it back in mentally because I knew both examples (CC edge and BS edge) were negative. But thanks to your proper example, I think I now see how to plug in a positive CC edge and get it to work right (I hope)...

Thanks again!

 

[Renzey]

I think I now see how to plug in a positive CC edge and get it to work right

If you're going to count only some of the time, then the time to count would be off the top of every new shoe. You could decide that you will for example, count until the first 1.5 decks are in the discard tray. That should have you counting around 35% of the time and just playing basic strategy 65% of the time.
But will that 65% where you're not counting and playing basic strategy really be basic strategy with flat bets?? Much of the time the answer will be NO! Instead, you'll be flat betting with BS half of those times, betting multiple units with an aggressive hand playing one-fourth of those times, and walking away from the shoe one-fourth of those times.

That's because those first 1.5 decks where you'll be counting will give you a preliminary "read" on each shoe. Remember that most shoes which go positive after the first 1.5 decks will remain positive, by and large. Most that go negative will hang in the negative area. Most that stay neutral won't usually get too far from neutral. Whatever the true count has become will become the new "norm", and things will tend to centralize themselves around that as the shoe progresses -- broadly speaking.

So if you only want to concentrate on the count a minority of the time, your time for concentrating will be at the beginning of each shoe. When the discard tray has reached your mark, stop counting, relax and play either a flat bet BS, a multiple unit "Plus Count Strategy", or walk away. I believe that would be the way to get the most out of "part time counting".

Here's an example. Say your count system is Hi/Lo. Say you also choose to count the first 1.5 decks only. When the discards reach their mark, if your RC is +7 (+1.5 TC) or higher, stop counting and go into multiple unit and advantage hand play mode for the rest of the shoe. If it's -6 (-1.3 TC) or worse, walk away. And if it's between +6 and -5, stay and just finish the shoe with flat bets and BS.

What is advantage hand play mode? It's BS plus all key index plays that you'd make at +1.5 TC or less. That would be:
Double with 9 vs. 2
Double with 11 vs. A
Stand with 16 vs. 10
Double with A/8 vs. 5 or 6

 

[QFIT]

If you're going to count only some of the time, then the time to count would be off the top of every new shoe.

This guarantees a terrible penetration in all shoes. You make your money toward the end of the shoe where there are more opportunities (assuming that you have counted all the way through). I'm not sure what counting part of the time means. But, it would seem that it would be better to count all the way through a shoe, or skip counting the shoe.

 

[Renzey]

This guarantees terrible penetration in all shoes. You make your money toward the end of the shoe where there are more opportunities. It would seem that it would be better to count all the way through a shoe, or skip counting the shoe.

Would it be better to count one-third of all your shoes all the way thru and just flat bet BS the other two-thirds of all your shoes -- or count the first one-third of every one of your shoes and either ramp up, stay flat or walk away based on your early count? Remember that if you thoroughly count just one-third of all your shoes, only about 20% of those hands will bear an advantageous count, amounting to maybe 7% of all your hands.

It's a viable question, but I vote for counting one-third the way into all your shoes, due to the central tendency of the true count moving forward beyond that point. Besides that, counting just a third of the way requires only about three different bet sizes and a half dozen index plays -- more suited to the recreational counter.

Remarks and critiques are welcome.

 

[QFIT]

The question hinges on your betting level when you are not counting. I would hope that the player would lower his bets when not counting. That would not be likely given your scenario. You are still assuming an edge, and therefore raising your bets, at some times during the non-counting period. That edge, overall, would be less than 0.1% given information from one-third penetration. This is a really high risk method of play.

I continue to maintain that basic strategy is superior, from a risk view, to poor counting techniques. Which is one of the reasons poor counters lose their bankrolls so quickly.

 

[Renzey]

You are still assuming an edge, and therefore raising your bets, at some times during the non-counting period. That edge, overall, would be less than 0.1% given information from one-third penetration.
.

Norm,
Notice that with this approach, you are not playing with only 1.5 decks of penetration, but indeed with full penetration. The hitch is that you are playing to full penetration with the moderately inaccurate assumption that the TC is what it was at the 1.5 dealt decks mark.

If you count thru those first 1.5 decks and then ramp up your bets because the RC was say, +9, you had a TC of +2 -- and that's what the TC will average thru the rest of the shoe. Yes, it will oscillate perhaps between zero and +4, but it will usually average around +2, and that's an advantage of about 0.6% with the appropriate half dozen index plays for a +2TC.

And yes, of course the recreational counter should bet 1 unit thru all shoes where the RC at the 1.5 deck mark was less than +7, but switch to maybe 3 units the rest of the way if that RC was +7 or 8, and and bet 5 or 6 units to the end if that early RC was +9 or higher. Also, at all 1.5 deck RC's of -6 or worse, his bets the rest of the way will be zero.

In effect, this player will be betting 1 unit all the way to the cut card thru shoes where the count is relatively neutral, he'll bet 3-to-6 units thru shoes that accumulated 7 or 8 or 10 extra high cards early on, and he'll bail on shoes that accumulated 6 or more extra low cards

 

[assume_R]

... the RC was say, +9, you had a TC of +2 -- and that's what the TC will average thru the rest of the shoe. Yes, it will oscillate perhaps between zero and +4, but it will usually average around +2

Renzey, consider also that while the TC theorem says that it will, on average, stay at +2, all the times when it's oscillating down towards +0 the player is overbetting, and all the times when it's oscillating up towards +4 the player is underbetting, severely affecting his RoR.

 

[Fun_at_21]

Thanks for the thoughts, from the both of you.

I might not be thinking clear in terms of math but couldn't one raise their bet early in a shoe (if even just a slight edge presented itself) but still just flat-bat that raised bet through the rest of the shoe at a long term edge (overall)? Because the only disadvantage you would've played at was right at the very beginning, and couldn't you MINIMUM bet, say, the first ten hands until you were able to see if the shoe turns positive at any point during those first hands? Wouldn't playing an advantageous shoe from a given starting point (with a raised but flat bet) outweigh having played a starting negative shoe at $5 for only the first handful of hands?

I know that there's no reason to believe that the shoes have to turn positive in which case you could lose numerous minimum bets from the top of several shoes thus making the losses harder to recoup once you do get a positive early shoe. Yet, it still seems that the odds of a shoe at least momentarily being positive during early hands would at least be fair (perhaps roughly half the time)?

In other words, if you always STARTED flat betting a shoe only when it was positive (say that 11 or 12 4's and 5's came out relative to one Ace in the first ten hands), then wouldn't you always have to end up with a long run advantage even if you played out the rest of the entire shoe?

I know in my example we can't ignore those FIRST hands. But if we are MINIMUM betting those and playing the majority of hands at an overall edge with a raised bet, would that not outweigh the times we are playing at a disadvantage when the shoe is fresh?

I'm sure my math and logic may be off somehwhere though...

 

[assume_R]

In other words, if you always STARTED flat betting a shoe only when it was positive, then wouldn't you always have to end up with a long run advantage even if you played out the rest of the entire shoe?

I know in my example we can't ignore those FIRST hands. But if we are MINIMUM betting those and playing the majority of hands at an overall edge with a raised bet, would that not outweigh the times we are playing at a disadvantage when the shoe is fresh?

A few hands at the beginning is fine to disregard.

Starting to bet only when the count is positive is called "wonging in".

The problem is if you have an advantage, start your "serious flat betting", then you stop counting, what if the count goes negative and you didn't know about it because you weren't using all the information available to you? (i.e. cards you saw but didn't feel like counting). The shoe might not stay positive the whole way to the end every time.

 

[Fun_at_21]

A few hands at the beginning is fine to disregard.

Starting to bet only when the count is positive is called "wonging in".

The problem is if you have an advantage, start your "serious flat betting", then you stop counting, what if the count goes negative and you didn't know about it because you weren't using all the information available to you? (i.e. cards you saw but didn't feel like counting). The shoe might not stay positive the whole way to the end every time.

Maybe that's where I'm confused because I'm assuming it WOULD stay positive as a WHOLE. I know that at POINTS within the shoe it will fluctuate but I'm looking at it in terms of the overall edge of a complete shoe based on the point it (hopefully) first turns positive (as a shoe from just that positive point forward).

For example, we know that we can safely flat-bet a normal 6 deck shoe playing good basic strategy and be guaranteed to lose .40-.50 of everything we will ever bet because we are always starting the shoe from beginning to end at an overall disadvantage (even though advantages and disadvantages will always shift during all the shoes we ever play). But in the end, if we always begin from a disadvantage, the fluctuations within shoes won't matter. Assuming we don't count and never raise bets to any known knowledge, we'll always expect to lose .40 - .50 of a percent on each fully shoe (or at least in the long run).

I'm just looking at the reverse of this from a positive perspective (assuming of course we can determine the shoe is positive sometime early on). For example, we could always just flat-bet any shoe to the very end if we always began this after all 5's were removed from the deck. We would never need to raise bets or lower them or keep running counts. We would know in the long run, we could flat-bet all the shoes from beginning to end and, overall, always be at an advantage.

I guess I'm wondering if early hands can't be minimum betted (and at least roughly counted) until we can (hopefully) see that the WHOLE shoe is currently at, say, a +.75 advantage. Then we could just raise that minimum bet ONCE and just flat bet that to the end with an overall edge of .75 (from the point it became .75 positive). Because you'd always be flat betting the big raised amount into a sudden favorable deck makeup while the initial .40 - .50 disadvantage was only getting $5 minimum bets put into it until the deck first turned positive (assuming it would half the time before the first deck or so went by).

I know that wonging would be the ideal way to do this because no disadvantage play has to be endured at the start. But sometimes at tables wonging in isn't "practical" or may just simply look more suspicious. I guess I was just wondering if, with a big enough jump from the minimum bet, one could still "wong in" so to speak while still playing (conservatively) from the very beginning of a shoe?

I'm not sure if that makes any more sense or not?

 

[assume_R]

I think you're logic is based on some flawed axioms. Specifically, in card counting, we are aiming to use the entire set of information at hand to make the best judgement about whether a hand is going to be "good" or "bad" when we bet, for the specific hand.

Let's say the count is +1 with a deck left. You know you have an advantage so you bet $100. Then, you get a blackjack (a 10 and an Ace). Now, after you see those 2 cards, the count drops to -1. You found out some more information, and hence make a new judgement about the next hand. For the next hand (count = -1), you now have a disadvantage, and hence bet less on the next round. This is all in the same shoe.

We are not trying to determine if we are "in a hot shoe", we are trying to determine, at each point of time through the shoe, how much to bet, and how to play the given hand.

Does this make sense to you?

 

[Fun_at_21]

I think you're logic is based on some flawed axioms. Specifically, in card counting, we are aiming to use the entire set of information at hand to make the best judgement about whether a hand is going to be "good" or "bad" when we bet, for the specific hand.

Let's say the count is +1 with a deck left. You know you have an advantage so you bet $100. Then, you get a blackjack (a 10 and an Ace). Now, after you see those 2 cards, the count drops to -1. You found out some more information, and hence make a new judgement about the next hand. For the next hand (count = -1), you now have a disadvantage, and hence bet less on the next round. This is all in the same shoe.

We are not trying to determine if we are "in a hot shoe", we are trying to determine, at each point of time through the shoe, how much to bet, and how to play the given hand.

Yes, it definitely makes sense. But I guess my curiosity was indeed more towards the "hot shoe" determination. I'm actually wondering if or how a flat-betting basic strategy approach might be maintained with only a minimum of early counting. I know that keeping running counts is the only precise way to best take advantage on specific hands but that is what I'm looking to bypass, if it can be largely done? At least for the majority of the shoe. Much like how those who wong in during an "advantage deck" don't necessarily need to raise or lower a bet to just keep that sudden overall deck advantage. If they flat bet, they'll just expect the deck to end with the advantage the deck carried from their joining in point (over the long run). Of course they would ALWAYS have to be wonging in when the deck was indeed a positive percentage (from when they play the first hand).

But I know that players who are wonging in are probably thinking in terms of positive counts, not positive decks (as a percentage).

But I was just curious as to what is (or isn't) hypothetically possible by just flat betting a one time raise with only minimal counting being done just off the top.

But maybe I'm answering my own questions and wonging in is the only real practical route to that...

 

[QFIT]

Norm,
Notice that with this approach, you are not playing with only 1.5 decks of penetration, but indeed with full penetration. The hitch is that you are playing to full penetration with the moderately inaccurate assumption that the TC is what it was at the 1.5 dealt decks mark.

Sorry, but this is incorrect. You ARE playing with 1.5 penetration and the assumption is anything but "moderately" inaccurate. If you stop counting after the first round, is that still "full penetration?" Suppose the dealer dealt to the last card. Are you saying that you are playing with 100% penetration?

 

[QFIT]

I ran three sims, 6D, S17, DAS, HiLo, Sweet16.

1. 75% penetration
2. 25% penetration
3. 75% penetration., but stop counting after 25% and assume the same TC through the remainder

SCORES
1. 15.4
2. -7.1
3. -0.8

In one sense, you are trying to “lock-in” the count after one-quarter of the shoe by halting further counting. But, to the same extent, you are also “locking-in” 25% penetration. One reason we look for deep penetration is not to play a lot of hands in a shoe, but to gain the increased accuracy of seeing a larger percentage of the cards. If we don’t count them, we don’t see them. More importantly, since count ranges are much narrower during the early part of the shoe, we lose high count opportunities. This is the main reason that penetration is important. (Explained in Modern Blackjack page 377.) These opportunities are thrown away by freezing the count.

Now, let’s look at the win rates. I have added sim 4 -- Basic strategy.
1. $9.38
2. -$2.27
3. -$1.00
4. -$2.83

Now, if we play BS for two hours and count for one hour, we make $9.83-2*-$2.83 or $3.72. But, if we use your method for three hours, we lose $3. The reason is that we always bet the low bet when playing BS. In your method, we bet according to the ramp even when not counting.

 

[assume_R]

fun_at_21, take some time to really look over the last response qfit posted and the sim results. No offense, but while you claim this to be a hypothetical situation, it sounds to me like you're trying to bypass the hard work required to become good and natural at counting. If it's hypothetical, I certainly apologize, but if it's laziness then I don't. Take the time to work hard and study, and the results will show.