Please Help: Conceptual understanding of knockout system

[lucas17]

I just started learning the knockout system from the book knockout blackjack and I know the system has a high betting correlation of .98. However I am having a hard time understanding intuitively how this systems gives you an advantage at all.

I understand that once you hit the key count, you start increasing your bet because that is when “you first have the advantage” as the book puts it. But how do you know you have the advantage at or above the key count? For example, the IRC for a six deck game is -20 and the key count is -4. Given that on average the RC after each deck dealt will increase by 4, you would expect to hit the key count in a six deck game after 4 decks have been dealt. So the only time you would have the advantage would be when you hit the key count before 4 decks are dealt, in which case there really is a high concentration of 10s and aces in the deck. But also will come the times when you don’t hit the key count until, say, you reach 4 1/2 decks in which case, according to the system you should increase your bet but the deck isn’t favorable to you because you are expected to be at an even higher count given the amount of deck that has been dealt.

So basically my question is this; how does the key count always ensure you have the advantage? Yes, if you hit the key count early in the shoe, before 4 decks are dealt, it makes sense you have the advantage and you should increase your bet. But won’t these advantageous situations be met with an equal amount of times when you hit the key count but don’t have the advantage because you’ve already dealt out 4 decks or more and are expected to hit that count anyways.

I hope I did a good job of explaining this, although there is a good chance I did not.

Thank you for the help!!!!!

 

[DSchles]

You did an excellent job of explaining your thoughts. Congratulations. This, of course, is one of the weaknesses of an unbalanced count, compared to a balanced, true-counted system. So, there's no such thing as a free lunch. You learn K-O for simplicity, to avoid having to estimate decks remaining and performing true count conversions, but at the expense of losing some accuracy in properly assessing your advantage at various levels of deck penetration.

I indeed, it matters quite a lot at what point you reach the Key Count. If you reached -4 (after starting at -20) after just 16 (all small) cards were dealt, that would represent a very different advantage situation than if you reached the Key Count at, say, 4 or 4.5 decks dealt. Reaching the Pivot, however, is a somewhat different phenomenon, as I suppose you understand, as well.

Don

 

[lucas17]

Thank you for the reply!

I’m trying to wrap my head around how the KO system can give you an edge when it comes to betting, especially since it touts a betting correlation of .98, higher than that of Hi lo which is .97.

It seems to me that all of the instances in which you hit the key count early in the shoe and increase your bets will be matched by an equal number of instances in which you hit the key count later than expected (past four decks dealt) and increase your bet in a bad situation. I don’t doubt the effectiveness of the system, I am simply trying to wrap my head around it.

Seperate question: if I play with at a six deck shoe, H17, Double any first two cards, DAS, 5 deck penetration, do you know the lowest spread I need to use to gin an advantage assuming I play the preferred KO system error free? $5000 bankroll. Do you think risk of ruin is too high? I know in the book it shows the expectation utilizing the system at different levels but for six deck it shows S17 which is not the case for the casinos near me.

Thank you!

 

[Dummy]

It seems to me that all of the instances in which you hit the key count early in the shoe and increase your bets will be matched by an equal number of instances in which you hit the key count later than expected (past four decks dealt) and increase your bet in a bad situation. I don’t doubt the effectiveness of the system, I am simply trying to wrap my head around it.

Both situations hurt you. KO uses the imbalance to predict how many decks have been played. It can be true counted by subtracting the imbalance for the number of decks seen from the change in the RC and dividing by the number of decks unseen. Doing this at the table is possible but sort of defeats why you use an unbalanced count to begin with, simplicity or no need for deck estimations. But you can get a feel for actual advantage by playing with some theoretical counts at home to get a TC for various levels of pen.

 

[lucas17]

Both situations hurt you. KO uses the imbalance to predict how many decks have been played. It can be true counted by subtracting the imbalance for the number of decks seen from the change in the RC and dividing by the number of decks unseen. Doing this at the table is possible but sort of defeats why you use an unbalanced count to begin with, simplicity or no need for deck estimations. But you can get a feel for actual advantage by playing with some theoretical counts at home to get a TC for various levels of pen.

But how does reaching a count of -4 (the key count) indicate an advantage? According to the KO system, that is when you raise your bet.

Thank you!

 

[lucas17]

Most likely based on an average of the EV at that point. Two things we should be aware of:
.
1.) At the Key Count (here on out denoted as KC), we are to assume that we have an EV at or around 0.00%. That means one below the KC, we are to experience -EV *on average*. The reason I bold *on average* is that there are certain shoes at any specific depth that offers us little to no advantage below the KC. This also means that one above the KC, we are to experience +EV, again, *on average*. Due to the unbalanced nature of KO, we are to assume that there are 6 small cards and 5 big cards. At the KC, there should be a balance of 5:5 for small:big cards. Once you go above the KC, you are running into +EV betting territory, according to the KO system.

2.) Now, the KC does not denote the correct -/+EV at that specific point. Let's take a single deck game. Our IRC is 0 and our KC is +4. At +5 (for example) with still a full deck remaining, we are to have around a 1% advantage (I forgot the specific EV value. Sorry, it has been a while since I used KO!) Now, let's say that we hit +5 again, BUT, with half a deck remaining! If you were to use the running count of +5 (here on out denoted as RC) at all deck levels, you would miss +EV situations. At one full deck, the true EV would be 1.0%, but at the half-deck level, the true EV would be about 2.0%! This is what is meant by "KO having you bet too much early into the shoe and too little later into the shoe." Another example, but with a shoe game. 6D game, rules won't apply here! You start out playing 1D out and your RC becomes 2. Quick! What is your true EV? The answer lies below:

EV_True = EV_RC / N_Cards * 52
​

That is, your true EV is the ratio of your RC and the number of cards remaining times the number of cards per deck. The above illustrates a method of "True Counting" the KO system and will help you bet less early on and betting more later on.


SO, to recap: The KC is the point at which you are close to breaking even with the house. Basically, a coin toss if you want to think of it like that. The KC tells us part of the picture of what the ratio of small cards to big cards is and why at different levels of a deck/shoe that the true EV for the current game is different even at the same KC.

Lemme know if I screwed up anywhere!

Thank you for the reply!

First, why is the key count the break even point where the expected value is 0.00% on average? Shouldn’t you EXPECT to get to a running count of -4 after 4 decks on average since every deck has a count of +4 on average? Reaching the key count could simply mean that 4 decks have been dealt out and the count has increased as expected on average, meaning the reaching of the key count would give you no information about an abnormally high concentration of tens and aces in the remainder of the deck. Also, why do you say the ratio of small to big cards is 5:5 at the KC?

I understand a true count conversion is possible for the KO system (by the way, you said the KC for single deck is 4 when according to the book it is 2, perhaps you meant the pivot point). My question is this: Without the true count conversion, how does the KO system identify good betting situations? For example, the KC for six deck game is -4. Once you reach -4, I don’t see how you have any information about the compsosition of the remainder of he deck. Why should you increase your bet at this arbitrary point? You should expect to reach this count since the RC for an unbalanced system increases naturally. Correct me if I’m wrong of course.

Again, thanks for the help, it means a lot.

 

[lucas17]

Before I go about answering further, let me find my copy of KOBJ.

Ok, sounds good

 

[lucas17]

Okay so, to answer some of your further questions:



Simply put, that is the point where the EV is closest to 0 and depending on what cards come next will determine if we will be betting in +EV territory or -EV territory.

But why is this? For balanced counts, the basis is that when the count increases more and more you begin to deviate more and more from normal deck composition and the concentration of tens and aces is higher than what is normal, which is beneficial to the player. But for an unbalanced count like the KO, the RC increases naturally and doesn’t give you any information about the composition of the deck (I’m not trying to sound like a know-it-all, I am clearly far from it. This is just how I understand it.) so how does the EV increase with the continued increase in the RC?


NOTHING, about the RC indicates the number of cards that have been played. The nature of a single deck to add up to 4 is simply due to the nature of the unbalanced count, as there are more llittle cards than big cards. Let me explain: Say you start off with a freshly shuffled 6D shoe. 16 cards come out, all of them a mix of either 2, 3, 4, 5, 6, or 7. After 16 of these cards come out, you are at your KC. Now, this is an extreme example, but one of *many* possible outcomes that can happen. So, what does this have to do with anything? Well, you mentioned that after 4 decks, we should be guaranteed on average to hit -4. This is false, as we are not guaranteed a god-damn thing in blackjack. You should not expect to hit -4 RC at the 2 deck level in a 6D game. Nor should you assume that at -4 RC, that 4 decks have been played. I illustrated an example in this paragraph earlier to counter your assumption. What you should be guaranteeing is that at or greater than -4, regardless of the numbers of decks that have been played, that there is when you should start raising your bets.

I am not saying that you are guaranteed to hit -4 after 4 decks dealt. I completely understand that you could reach it after 16 cards. But there are more small cards than big cards being counted, so the count will naturally rise, so a high count indicates no deviation from normal deck composition and no advantage ( Again, how I am thinking about it. I don’t even believe what I just said is true, just trying to understand me how it’s false.)



Sorry, that should be for the Pivot Point. Not the Key Count. Disregard!



Please see the bold part!

Here in lies the issue you may not be seeing. At this point, the system developers tested rigorously that at the -4 RC point in a 6D game, is where you will traverse from -EV betting territory to +EV betting territory. As for me, the remainder of your question is simply noise. What I am reading is, "The nature of the underlying system makes no intuitive sense." Which, for an unbalanced system, makes...well...sense! I am going to venture that hours of simulating 6D games by both Ken and Olaf that they came to the answer that the KC is -4. Now, this is also based on the running count nature of the system, as a RC at the top of a fresh shoe is different in EV than that of the same RC at the last two decks left of the shoe.

I am definitely not trying to deny that the system is effective, nor am I denying the mathematical results of simulation. It just doesn’t make intuitive sense. And there may be no intuitive explanation, as you hinted at above.


Composition of a shoe is of no importance for the system. The system is built to be a simple linear approximation of the EV for all levels of a deck/shoe. If you want to true count (TC) the KO system, then I would recommend reKO, or even better, High-Low! KO was not built to be an academic ponder. Rather a simple system that gains +EV for you in the long run. The system has been mathematically/systematically tested by mathematicians. It is a sound system. The data in the book, I reckon, is valid.

I guess my question is how your EV breaks even and begins to increase if there is no change of composition in the deck on average at the KC.

Again, the reason to true count KO is to correctly size your bets depending on how deep you are in the shoe. A +6 RC at 5 decks has a lower +EV than a +6 RC at 2 decks. The first is a +1.2 TC and the last is a +3 TC. This, to me, suggests that your advantage is .6% and 1.5%, assuming .5% per TC.


I take it you are a beginner counter? No worries, these things will come to make sense. For now, just understand that when both K&O suggest ramping your bets at -4 to +6 in a 6D shoe, you should. Disregard the number of decks that have been played unless you want to true count the game. And, again, if you want to TC the game, I would highly recommend High-Low as a great starter system for any player.

I am a beginner. I started with the Hi Opt 1 but wasn’t a fan of the multi parameter aspect as my first system. Regarding the KO, I understand all of the simulation and data in the book and don’t deny that any of it is valid, I just am trying to know exactly why i gain an advantage at KC and everything else mentioned above.

Thanks a million for the help.

 

[lucas17]

Okay so, to answer some of your further questions:



Simply put, that is the point where the EV is closest to 0 and depending on what cards come next will determine if we will be betting in +EV territory or -EV territory.



NOTHING, about the RC indicates the number of cards that have been played. The nature of a single deck to add up to 4 is simply due to the nature of the unbalanced count, as there are more llittle cards than big cards. Let me explain: Say you start off with a freshly shuffled 6D shoe. 16 cards come out, all of them a mix of either 2, 3, 4, 5, 6, or 7. After 16 of these cards come out, you are at your KC. Now, this is an extreme example, but one of *many* possible outcomes that can happen. So, what does this have to do with anything? Well, you mentioned that after 4 decks, we should be guaranteed on average to hit -4. This is false, as we are not guaranteed a god-damn thing in blackjack. You should not expect to hit -4 RC at the 2 deck level in a 6D game. Nor should you assume that at -4 RC, that 4 decks have been played. I illustrated an example in this paragraph earlier to counter your assumption. What you should be guaranteeing is that at or greater than -4, regardless of the numbers of decks that have been played, that there is when you should start raising your bets.



Sorry, that should be for the Pivot Point. Not the Key Count. Disregard!



Please see the bold part!

Here in lies the issue you may not be seeing. At this point, the system developers tested rigorously that at the -4 RC point in a 6D game, is where you will traverse from -EV betting territory to +EV betting territory. As for me, the remainder of your question is simply noise. What I am reading is, "The nature of the underlying system makes no intuitive sense." Which, for an unbalanced system, makes...well...sense! I am going to venture that hours of simulating 6D games by both Ken and Olaf that they came to the answer that the KC is -4. Now, this is also based on the running count nature of the system, as a RC at the top of a fresh shoe is different in EV than that of the same RC at the last two decks left of the shoe.

Composition of a shoe is of no importance for the system. The system is built to be a simple linear approximation of the EV for all levels of a deck/shoe. If you want to true count (TC) the KO system, then I would recommend reKO, or even better, High-Low! KO was not built to be an academic ponder. Rather a simple system that gains +EV for you in the long run. The system has been mathematically/systematically tested by mathematicians. It is a sound system. The data in the book, I reckon, is valid.

Again, the reason to true count KO is to correctly size your bets depending on how deep you are in the shoe. A +6 RC at 5 decks has a lower +EV than a +6 RC at 2 decks. The first is a +1.2 TC and the last is a +3 TC. This, to me, suggests that your advantage is .6% and 1.5%, assuming .5% per TC.


I take it you are a beginner counter? No worries, these things will come to make sense. For now, just understand that when both K&O suggest ramping your bets at -4 to +6 in a 6D shoe, you should. Disregard the number of decks that have been played unless you want to true count the game. And, again, if you want to TC the game, I would highly recommend High-Low as a great starter system for any player.

Well my reply got a little messed up, but my responses are after yours within the quote of your entire response.

 

[KewlJ]

I would highly recommend High-Low as a great starter system for any player.

Really? "starter system"?

 

[London Colin]

At the risk of muddying the waters still further, here's an overview of what I (think) I understand about all this. (My knowledge of unbalanced counts mainly comes from reading the section on the red 7 count in Snyder's Blackbelt in Blackjack.)

1. The main purpose of using an unbalanced count is to simplify the task of counting by getting rid of the need to divide the RC by the number of remaining decks.
2. This simplification is, perhaps inevitably, going to lead to situations in which you either think you have the advantage when you do not or, conversely, don't think you have the advantage when you do.
3. When choosing the RC at which to raise your bets, the trick must therefore be to find the point at which the gains from correctly raising your bets outweigh the losses from incorrectly raising them. (This is presumably best found by simulation.)
4. In a count such as red 7, the imbalance per deck is +2, making the pivot the ideal point to raise your bet (around 1% above the off-the-top house edge, so typically a 0.5% player advantage. But with the +4 imbalance of KO it would be too wasteful to wait for a +1.5% advantage at the pivot, so you have to raise your bets earlier. Doing this at a RC of -4 means that, after around 4 decks of 6-deck shoe have been dealt, the KO system is directing you to raise your bets before you have an advantage, and this is part of the price you pay for the simplicity of basing your decisions purely on the RC.

If I've got all that about right, then a couple of thoughts which then come to mind are -

1. The penetration must make a difference when comparing the efficiency of KO with that of other (particularly, balanced) counts. After all, if 2 decks are cutoff then the problems of what happens after 4 decks have been dealt are moot.)
2. Could it be that the mysterious phenomenon known as the 'floating advantage' is helping to lift some of the late-in-the shoe borderline decisions into +EV territory?

 

[DSchles]

At the risk of muddying the waters still further, here's an overview of what I (think) I understand about all this. (My knowledge of unbalanced counts mainly comes from reading the section on the red 7 count in Snyder's Blackbelt in Blackjack.)

1. The main purpose of using an unbalanced count is to simplify the task of counting by getting rid of the need to divide the RC by the number of remaining decks.
2. This simplification is, perhaps inevitably, going to lead to situations in which you either think you have the advantage when you do not or, conversely, don't think you have the advantage when you do.
3. When choosing the RC at which to raise your bets, the trick must therefore be to find the point at which the gains from correctly raising your bets outweigh the losses from incorrectly raising them. (This is presumably best found by simulation.)
4. In a count such as red 7, the imbalance per deck is +2, making the pivot the ideal point to raise your bet (around 1% above the off-the-top house edge, so typically a 0.5% player advantage. But with the +4 imbalance of KO it would be too wasteful to wait for a +1.5% advantage at the pivot, so you have to raise your bets earlier. Doing this at a RC of -4 means that, after around 4 decks of 6-deck shoe have been dealt, the KO system is directing you to raise your bets before you have an advantage, and this is part of the price you pay for the simplicity of basing your decisions purely on the RC.

If I've got all that about right, then a couple of thoughts which then come to mind are -

1. The penetration must make a difference when comparing the efficiency of KO with that of other (particularly, balanced) counts. After all, if 2 decks are cutoff then the problems of what happens after 4 decks have been dealt are moot.)
2. Could it be that the mysterious phenomenon known as the 'floating advantage' is helping to lift some of the late-in-the shoe borderline decisions into +EV territory?

FWIW, I agree with everything you've written, including the very last point. But, as you probably know from the study, there's not much going on, FA-wise at the 4/6 level. All of the good stuff starts much later, at depths that we rarely if ever see these days.

Nice post!

Don