how many and how long

how many hands/ rounds would it take to actually see a true house edge on blackjack?

how long in real life does it take to complete a round?

do you think anyone actually sees that many hands/rounds in their "lifetime"?

if they dont see that many, why are we told we must play blackjack for the long run and not the "short run"

i am tired of not hitting on certain situations and the dealer getting my cards. and i do hit, but i just am wondering these things...

We know that eating kale is better for us than eating double cheeseburgers. It doesn't make it easier to accept.
Same thing. No one tells you you must play any hand any way. They merely point out the best option. If you chose to play sub optimally, you may go broke before the Long run.
You can play the mathematically correct way, or any way you want.

I think about this a lot. For example, splitting 8s is so I lose less in the long run. But maybe I don't want to lose two bets now?

One day, I won't split 8s and then quit playing blackjack forever!

Quote: smoothgrh

I think about this a lot. For example, splitting 8s is so I lose less in the long run. But maybe I don't want to lose two bets now?

One day, I won't split 8s and then quit playing blackjack forever!

Assuming you don't spontaneously explode.

Quote: heatmap

how many hands/ rounds would it take to actually see a true house edge on blackjack?

One. Only way I can see myself being wrong here is if there is an example of an 'untrue' house edge

Quote: heatmap

if they dont see that many, why are we told we must play blackjack for the long run and not the "short run"

For "long" profits, rather than "short" profits -- assuming that we're doing it right and the house edge is negative. If the house edge is positive, playing less, or "shorter", is better than playing "longer".

The one thing about most blackjack is that the house edge changes based on what cards have been played and what remains. A common solution that I like is to vary bets based on whether the player edge is positive or negative. Perhaps an even better solution is the ZK method of only playing when the player has the edge.

Quote: TomG

Perhaps an even better solution is the ZK method of only playing when the player has the edge.

And when no Chinese prison decks are in play. Never on the weekends, never on holidays and never on the strip. I'm sure he has a laundry list of nevers but I don't think there's enough room here to post them all.

Quote: AxelWolf

And when no Chinese prison decks are in play. Never on the weekends, never on holidays and never on the strip. I'm sure he has a laundry list of nevers but I don't think there's enough room here to post them all.

Never play when your shoe is empty.

Quote: AxelWolf

And when no Chinese prison decks are in play. Never on the weekends, never on holidays and never on the strip. I'm sure he has a laundry list of nevers but I don't think there's enough room here to post them all.

No debt. No alarm clocks.

Quote: heatmap

how many hands/ rounds would it take to actually see a true house edge on blackjack?

how long in real life does it take to complete a round?

do you think anyone actually sees that many hands/rounds in their "lifetime"?

if they dont see that many, why are we told we must play blackjack for the long run and not the "short run"

i am tired of not hitting on certain situations and the dealer getting my cards. and i do hit, but i just am wondering these things...

1. The maths

This is not about long term or short term.

This is about the law of large numbers.
The casino works on a large number of hands. The player not.
Let N be the number of rounds, E be the (negative) expected value of a round (rather than house edge, which is relative to wager) and S be the standard deviation. The total value is distributed around N*E with a std dev of sqroot(N)*S. Then with N large enough (50 or 100, for instance)The probability of total value being higher than 1.96 std devs above the average is 5% (Normal-Gauss distribution).

So, when NE+1,96 sqrt(N)S is zero, that is N=(1,96S/-E)^2, the total has 95% chance of being negative.
For instance, with E=-1% and S=4, this gives N=784^2= 614656 rounds, for an Expected - 6146,56 bet units and a SD of 3136 units.

As for your question, if I understand what you mean, the probability of falling exactly in a small neighboring of the average is close to zero and goes closer with N increasing. But the probability of falling in a *relative* neighboring, I.e. close to the house edge, increases and tends to 100% in the limit. That limit is in the order of several million hands.

2. How to play?

Most mathematical people tend to think (and impose) that there is only one way to play, and that is optimizing Expected Value. This is a mathematical concept, and in order to ease the explanation they use the notion of long term. But that is NOT correct, especially if you play a negative expectation game. The reason is they use a statistical concept for an individuel situation.

Read any book on investment or portfolio management, any book on game theory or decision in uncertainty, and you’ll see nobody else thinks like that.
Variance should be taken into account. Bankroll management also. Time. And the objective (so-called Utility) of the decider.

In essence, you do what you feel best, on the condition that you know what you are doing.

If you play 100 hands, the square root would be 10 hands; HA would be 0.5 hands.^^^^^^^^^(-10.5, +9.5)
If you play 400 hands, the square root would be 20 hands; HA would be 2 hands.^^^^^^^^^^(-22, +18)
If you play 900 hands, the square root would be 30 hands; HA would be 4.5 hands.^^^^^^^^^(-34.5, +25.5)
If you play 1,600 hands, the square root would be 40 hands; HA would be 8 hands.^^^^^^^^^(-48, +32)
If you play 2,500 hands, the square root would be 50 hands; HA would be 12.5 hands.^^^^^^^(-62.5, +37.5) *
If you play 3,600 hands, the square root would be 60 hands; HA would be 18 hands.^^^^^^^^(-78, +42)
If you play 4,900 hands, the square root would be 70 hands; HA would be 24.5 hands.^^^^^^^(-94.5, +45.5)
If you play 6,400 hands, the square root would be 80 hands; HA would be 32 hands.^^^^^^^^(-112, +48)
If you play 8,100 hands, the square root would be 90 hands; HA would be 40.5 hands.^^^^^^^(-130.5, +49.5)

If you play 10,000 hands, the square root would be 100 hands; HA would be 50 hands.^^^^^^^(-150, +50) PEAK
If you play 40,000 hands, the square root would be 200 hands; HA would be 200 hands.^^^^^^(-400, 0)
If you play 90,000 hands, the square root would be 300 hands; HA would be 450 hands.^^^^^^(-750, -150)
If you play 160,000 hands, the square root would be 400 hands; HA would be 800 hands.^^^^^(-1200, -400)
If you play 250,000 hands, the square root would be 500 hands; HA would be 1,250 hands.^^^^(-1750, -750)
If you play 360,000 hands, the square root would be 600 hands; HA would be 1,800 hands.^^^^(-2400, -1200)
If you play 490,000 hands, the square root would be 700 hands; HA would be 2,450 hands.^^^^(-3150, -1750)
If you play 640,000 hands, the square root would be 800 hands; HA would be 3,200 hands.^^^^(-4000, -2400)
If you play 810,000 hands, the square root would be 900 hands; HA would be 4,050 hands.^^^^(-4950, -3150)
If you play 1,000,000 hands, the square root would be 1,000 hands; HA would be 5,000 hands.^^(-6000, -4000)

The more you play at a 0.5% HA, the more it's going to eat up any positive variation you can get until you are totally in the hole. Win early, and win big!

Seems there's a peak somewhere between 10,000 and 40,000 hands. Let me see if I can find it.
If you play 12,100 hands, the square root would be 110 hands; HA would be 60.5 hands.^^^^^^(-170.5, +49.5)
If you play 15,625 hands, the square root would be 125 hands; HA would be 78 hands.^^^^^^^(-203, +47)
If you play 22,500 hands, the square root would be 150 hands; HA would be 112.5 hands.^^^^^(-262.5, +37.5) *
If you play 30,625 hands, the square root would be 175 hands; HA would be 153 hands.^^^^^^(-328, +22)

* If I wanted a win goal of 37.5 hands, I should settle for it somewhere between hands 2,500 and 22,500, and the sooner the better. I could lose 7 times 37.5 units at the latter number of hands.

I'd probably stick to near 400 hand sessions and either lose 25 hands or win 20 hands and quadruple my buy-in.

These calculations are for 1 standard deviation. YMMV
If I play 400 hand sessions and have a 2 standard deviation streak, I could be down 42 hands or up 38 hands, (and get to my 37.5 unit win goal).
If I play 10,000 hand sessions and have a 3 standard deviation streak, I could be down 350 hands or up 250 hands.

Quote: heatmap

how many hands/ rounds would it take to actually see a true house edge on blackjack?

~50k hands on average.

Quote: heatmap

how long in real life does it take to complete a round?

One "round" of blackjack? Probably 30-45 seconds. Complete a round of real life? On average 71 years of age... thanks google.

Quote: heatmap

do you think anyone actually sees that many hands/rounds in their "lifetime"?

EASILY. I haven't played blackjack in like 5 years and I've still seen hundreds of thousands of hands, easily.

Quote: heatmap

if they dont see that many, why are we told we must play blackjack for the long run and not the "short run"

They will see that many, first off. Secondly I'm giving 3SD worth of confidence. If someone simply played 100 hours of blackjack in their LIFETIME, so like 2 hours per year after they hit 21, and they flat bet $10 at 80 hands per hour on an AVERAGE game of blackjack:

OriginalSD = 1.15 * 10 = 11.5

EV(8000 hands per life) = (8000*10)*(-.005) = -$400
SD(8000 hands per life) = Sqrt(8000) * 11.5 = $1028.59... 3SD = 3085.77

So in the persons life, with the above conditions, ON AVERAGE the average person will LOSE $400 +/- $3085.77.

Again though, for the MASS MAJORITY of people that "play blackjack" even for fun they WILL see enough hands to hit the long run in their life. An agerage person playing 2 hours per WEEK for 50 years with the same conditions above would look like this:

5200 hours, 80 hands per hour = 416,000 hands

EV(416,000 hands) = (416,000*10)*(-.005) = -$20,800
SD(416,000 hands) = Sqrt(416,000) * 11.5 = 7417.28... 3SD = 22,251.83

So a "real" average person playing 2 hours per week will on average LOSE $20,800 +/- $22,251.83. So if they were in the <1% of the LUCKIEST PLAYERS ON EARTH they would break even. EVERYONE ELSE IS LOSING MONEY.

Also, the casino isn't worried about 1 person seeing the long run. They're worried about the law of large numbers. To the casino the patron doesn't matter, the hands are what matters... Similarly for card counters the casino they count at doesn't matter, it's the number of hands they get that matters. The casino gets 50,000 hands of action a day (probably way more) so the casino is seeing a long run on a daily basis, and thus making guaranteed profits on a daily basis, for the most part.

Quote: heatmap

i am tired of not hitting on certain situations and the dealer getting my cards. and i do hit, but i just am wondering these things...

Basic strategy doesn't care what you feel, or what you like, or what you don't like... It is the mathematical absolute best way to play the game with the lowest house edge possible for the given rule set. That's the great thing about math, it's indifferent. The right play is the right play, and that's all that there is to be said about it.

I love you all

Quote: Romes

That's the great thing about math, it's indifferent. The right play is the right play, and that's all that there is to be said about it.

Math is a dish best served cold.

I'm gonna say the peak spread is at 10,000 hands, and the downside gets worse above that, and the upside trends down above that. Throw in a bit of luck and you could be up 50 hands before or after that, or never.

Quote:

Basic strategy doesn't care what you feel, or what you like, or what you don't like... It is the mathematical absolute best way to play the game with the lowest house edge possible for the given rule set. That's the great thing about math, it's indifferent. The right play is the right play, and that's all that there is to be said about it.

[My emphasis]

The first assertion is incorrect: basic strategy, as the name ‘basic’ suggests, is just a good approximate of the ‘absolute’ best way to optimize the H.E. criterion.
The absolute best is more complicated, and depends on an exact count. Even current counting strategies are only an approximate. The aim is not to do the ‘best’, it is to be positive.

The second assertion is incorrect. It confuses  aiming at lowest house edge with « right play ». What is right play, anyway? Like I said above, decision theory has shown that this question has no objective answer. All depends on individual aims.

Quote: ChumpChange

I'm gonna say the peak spread is at 10,000 hands, and the downside gets worse above that, and the upside trends down above that. Throw in a bit of luck and you could be up 50 hands before or after that, or never.

What do you call ‘peak spread’?

If it is the largest spread (in terms of confidence interval), then it is never reached. The more hands, the more spread.

If it is the maximum upper bound of said interval, it comes at 1/4 of the ‘negative total’ N referred to in my above post. Or

N = (zS/2E)^2

where z is the standard Normal number of deviations needed for your desired probability : 1.960 for 95% — 2.326 for 99% — 2.576 for 99,5% — 3.090 for 99,9%

This is the number of hands N for which you reach the highest total obtainable at a reasonable probability (5, 1, 0.5 or 0.1 % in the examples). In this case, you want to use a conservative level.
For Blackjack, E=0,5% and S=1.14
With z=2, we get N=228^2= 51,984 hands. The 2S conf. interval is then -259.92 +/- 519.84 units. The upper bound is … 259.92 units.
Your figure of 10 000 hands is reached with a z = 0.8772, corresponding to a confidence level of about 81.3%. I don’t think that was what you meant.

Quote: heatmap

how many hands/ rounds would it take to actually see a true house edge on blackjack?

how long in real life does it take to complete a round?

do you think anyone actually sees that many hands/rounds in their "lifetime"?

if they dont see that many, why are we told we must play blackjack for the long run and not the "short run"

i am tired of not hitting on certain situations and the dealer getting my cards. and i do hit, but i just am wondering these things...

Please note that Heatmap asked two different ‘families’ of questions.

One is the ‘how many, how long’ family. The other is ‘why should I follow basic strategy?’

My calculations concern the first family. From an external point of view or, if you prefer, from the casino’s point of view: all hands are undifferentiated, their total is considered beforehand.
From a player’s p.o.v., of course, there is a sequential structure. As soon as he/she has won or lost the first hand, all calculations are to be redone. It makes no sense to define a number of hands to play in advance.

Remember your own experience : as time goes on playing, as your bankroll evolves, your objectives change. So do the distribution of final amounts won/lost. Only a robot (and maybe an AP? 😉) plays in a risk-neutral, bankroll-independent way — that is, basic strategy....

You want to throw more confidence into the equation, it looks like the peak of the curve occurs at a higher number of hands. Thanks. I'm just using some worn out math from 25 years ago. It wasn't meant to be 99%, but probably more like 68.2% around a zero point with an HA offset.

Who says leaving sessions a winner means nothing? People who are unsuccessful gamblers, that's who!

I'd go so far as to say the majority of card counters are gamblers who could never win in their whole lives and then turned to something else.

Those who were winning before and then turned ADDITIONALLY to card counting, those are the ones to watch out for!

I just ran across this article this morning. It's interesting how the bet sizes go way up on a count of +2 or more with a 1% to 4%+ advantage over the house. I'd try to incorporate that into a winning streak progression, like when the dealer keeps busting time after time.
A to Z Counting Cards in Blackjack - Articles - Wizard of Vegas https://wizardofvegas.com/articles/A-to-Z-Counting-Cards-in-Blackjack/

Earlier today: In between waiting for a new Bacc. shoe to shuffle I went up to a six deck BJ shoe and watched the count get to +12, while the only other player on there was playing and I a mere spectator. I asked the other player if he minded if I jumped in, he said sure - he had been steadily LOSING as the count was getting more positive, and had been decreasing his bet actually. I tossed a grand on there (min. black chip table), got a 6/4 for a ten, dealer face card up. No double of course. I hit got a 2, hit again got a 7 for a 19, dealer turns over a 5, busts with a ten. Anyway, yes, the count "worked" for that one, at least for me.

If this is your last hand of blackjack ever ... You should still split 88 vs T.

If the true count was +12, which it wasn't, you'd be a fool not to double.