Ploppy would do whatever I want him to do

21gunsalute

Well-Known Member
#61
assume_R said:
The TC's in your examples changed, yes. Nobody is arguing that. But would you be able to predict how it would have changed? No, because on average it has as much a chance of decreasing as it does of increasing. That's all the TCT says. On average it stays the same. Not that you should predict it to remain the same. But that it won't tend towards one way or another.
That's exactly my point. The TCT tells you nothing of any value or usefullness. All it gives you is an expected value which is not likely to reflect any actual cirmustances.
 
#63
This discussion was about the player before you taking one card. The people who said that it was a high count so he is going to take my ten are wrong. The point is every card has the same likelihood to a high card or low card. It is not until the card is seen that you know the cards effect on the TC. If it is a high card your chances the next card is a ten go down a hair but if it is a low card your chances of drawing a ten go up more than they go down in the other case. These 2 events balance each other perfectly in frequency times effect showing the true count theorem applies in this situation. The other player doubling has no effect on your play. If you were playing in a store where the double card is dealt face down the situation would be the same. I doubt people would be arguing he is going to take their ten. They would look even more foolish than they already do.
 

MangoJ

Well-Known Member
#64
Whether or not TCT is useful on any given situation, it is never a bad idea to know as much as possible about the game you're playing. I see it as a long-term investment, since you never know what kind of other game will come across in the future, where TCT (or other theorems) will come handy.
 

aslan

Well-Known Member
#65
21gunsalute said:
That's exactly my point. The TCT tells you nothing of any value or usefullness. All it gives you is an expected value which is not likely to reflect any actual cirmustances.
I have been translating "expected" to mean "probable". If that is true, why wouldn't it "likely" reflect actual circumstances? Or am I still missing the point?
 

MangoJ

Well-Known Member
#66
aslan said:
I have been translating "expected" to mean "probable". If that is true, why wouldn't it "likely" reflect actual circumstances? Or am I still missing the point?
"Expected" doesnt have much to do with "probable" or even "most probable".
Example:
On a fair coin flip game of a billion rounds paying even money, the most probable outcomes are well ahead or well behind. However, being behind and being ahead has same probability (since the game is symmetric to your opponent's side). That means the expected win of that game is zero. Of course only in rare cases you come out at exactly zero after a billion rounds.
 

21gunsalute

Well-Known Member
#67
aslan said:
I have been translating "expected" to mean "probable". If that is true, why wouldn't it "likely" reflect actual circumstances? Or am I still missing the point?
Do an experiment and True Count several shoes, recording or remembering the TC at various intervals such as a deck or 2 decks. According the the TCT, the expected TC will be the same at each interval. In actual practice though the TC rarely remains constant. You could probably even use the strategy trainer here for such an experiment.
 

aslan

Well-Known Member
#68
MangoJ said:
"Expected" doesnt have much to do with "probable" or even "most probable".
Example:
On a fair coin flip game of a billion rounds paying even money, the most probable outcomes are well ahead or well behind. However, being behind and being ahead has same probability (since the game is symmetric to your opponent's side). That means the expected win of that game is zero. Of course only in rare cases you come out at exactly zero after a billion rounds.
So expected is more like the middle ground or average, but you can expect the actual case to usually fall somewhere on one side or the other. Then, the only probably is that you are probably not "far" off, lol, but that does not give a whole lot of comfort-- still, it's better than nothing. Anyway, am I getting warm?
 

MangoJ

Well-Known Member
#69
aslan said:
So expected is more like the middle ground or average, but you can expect the actual case to usually fall somewhere on one side or the other. Then, the only probably is that you are probably not "far" off, lol, but that does not give a whole lot of comfort-- still, it's better than nothing. Anyway, am I getting warm?
Yes, expected value is commonly referred to as "average". They are not the same in statistical terms (the expected value of the average is the expected value, but the average is not the expected value). But to keep it simple - yes if you average over a "large enough" set, the average will be the expected value.

Your next question "what is large enough" is valid, it depends on the variance of the quantity you are studying. For high quantities, you need a larger set to get the average reasonable close to the expected value.
 

aslan

Well-Known Member
#70
MangoJ said:
Yes, expected value is commonly referred to as "average". They are not the same in statistical terms (the expected value of the average is the expected value, but the average is not the expected value). But to keep it simple - yes if you average over a "large enough" set, the average will be the expected value.

Your next question "what is large enough" is valid, it depends on the variance of the quantity you are studying. For high quantities, you need a larger set to get the average reasonable close to the expected value.
So when you say "expected" it ironically is not that you expect this to be the actual value, or even very close to it, but that it is the best you can do considering the wide range of possibilities. You can probably surmise that the actual value will likely be closer to the expected value than to either extreme. Can you say any more than that? So when we pick the strategy with the highest EV, we are actually picking our best educated guess as to what the actual result will be.
 

MangoJ

Well-Known Member
#72
aslan said:
So when you say "expected" it ironically is not that you expect this to be the actual value, or even very close to it, but that it is the best you can do considering the wide range of possibilities. You can probably surmise that the actual value will likely be closer to the expected value than to either extreme. Can you say any more than that? So when we pick the strategy with the highest EV, we are actually picking our best educated guess as to what the actual result will be.
You are not trying to predict or guess any specific result. That is what gamblers do. "You missed your winning result ? Oh no problem, let's try again, eventually we must hit the win - right ?"

What APs do is something different, which is enjoying the "central limit theorem" (call it long run): If you average a random quantity over a large set, your result will be reasonable close to the expectation value of that quantity. You can even compute how likely you are how close. Nothing more, nothing less. If you play a positive expectation strategy long enough, you are expected to be close to that value.

Statistics is never about the single value of an individual observation. It is about this beautiful property - which, although all individual observations are intrinsically random, - guarantees you the expectation value "almost surely".
I don't find it ironical, maybe you don't agree on the wording "expectation" with literature. You have all right to disagree with the wording. You can claim you don't live for the central limit and this is perfectly right. But that doesn't alter statistical laws within that limit.
 

aslan

Well-Known Member
#73
MangoJ said:
You are not trying to predict or guess any specific result. That is what gamblers do. "You missed your winning result ? Oh no problem, let's try again, eventually we must hit the win - right ?"

What APs do is something different, which is enjoying the "central limit theorem" (call it long run): If you average a random quantity over a large set, your result will be reasonable close to the expectation value of that quantity. You can even compute how likely you are how close. Nothing more, nothing less. If you play a positive expectation strategy long enough, you are expected to be close to that value.

Statistics is never about the single value of an individual observation. It is about this beautiful property - which, although all individual observations are intrinsically random, - guarantees you the expectation value "almost surely".
I don't find it ironical, maybe you don't agree on the wording "expectation" with literature. You have all right to disagree with the wording. You can claim you don't live for the central limit and this is perfectly right. But that doesn't alter statistical laws within that limit.
Expected value over the long run I can live with, and I see now that that's how it is meant, but expected value in the short run, well, that is what seems (or seemed) to me a bit ironical. OF course now I see the context in which it is used, so it is not ironical at all. If that makes any sense:confused: :laugh:
 
Top