Equation
I talked with the resident math wizard and he came up with this (I'll call it the "Bertsos Bonus Solution"):
Lets say you put in Y dollars you get a bonus of B dollars
and you have to risk R dollars to get out. Also say your chance of
winning a particular bet is p, losing is q=1-p.
If you double up each time starting with Y+B after n bets you have bet
an amount of (2^n-1)*(Y+B). So first find the biggest n such that
this is still less than R.
So what you are doing is making n bets (or until you lose) and
then making a last bet of R-(2^n-1)*(Y+B) to top off your risk.
The only way you get money back is if you win all n bets
prob=p^n . If you win the top off bet it turns out you have
(2^n)*(Y+B) +top off =
(2^n)*(Y+B) + R-(2^n-1)*(Y+B)
prob is (p^n)*p
If you lose the top off you have
(2^n)*(Y+B) - R-(2^n-1)*(Y+B) prob is (p^n)*q
So our average amount left is (after simplifying)
(p^n) * ( (2^n)*(Y+b) + (p-q)*(R-(2^n-1)*(Y+B) )
subtract your original Y from that and you have
your average return.
Also if you want to find n in one line I would use
n = int (log(R/(Y+B) +1)/log2)
Example:
Y=100, B=100, p=.45, R=500.
So you
bet 200 - have 400
top off risk, bet 300 you have either 100 or 700 if you lose or win.
prob 100= .45*.55=0.2475
prob 700= .45*.45=0.2025
average= 100*.2475+700*.2025=166.5
So your average return is 166.5-Y=66.5
I talked with the resident math wizard and he came up with this (I'll call it the "Bertsos Bonus Solution"):
Lets say you put in Y dollars you get a bonus of B dollars
and you have to risk R dollars to get out. Also say your chance of
winning a particular bet is p, losing is q=1-p.
If you double up each time starting with Y+B after n bets you have bet
an amount of (2^n-1)*(Y+B). So first find the biggest n such that
this is still less than R.
So what you are doing is making n bets (or until you lose) and
then making a last bet of R-(2^n-1)*(Y+B) to top off your risk.
The only way you get money back is if you win all n bets
prob=p^n . If you win the top off bet it turns out you have
(2^n)*(Y+B) +top off =
(2^n)*(Y+B) + R-(2^n-1)*(Y+B)
prob is (p^n)*p
If you lose the top off you have
(2^n)*(Y+B) - R-(2^n-1)*(Y+B) prob is (p^n)*q
So our average amount left is (after simplifying)
(p^n) * ( (2^n)*(Y+b) + (p-q)*(R-(2^n-1)*(Y+B) )
subtract your original Y from that and you have
your average return.
Also if you want to find n in one line I would use
n = int (log(R/(Y+B) +1)/log2)
Example:
Y=100, B=100, p=.45, R=500.
So you
bet 200 - have 400
top off risk, bet 300 you have either 100 or 700 if you lose or win.
prob 100= .45*.55=0.2475
prob 700= .45*.45=0.2025
average= 100*.2475+700*.2025=166.5
So your average return is 166.5-Y=66.5
my problem with the whole thing is you could give me a 500 % bonus
i still have to win a hand
if i bet $1 it seems i have 48-52 chance of winning vs losing a single hand
if i bet $5 40-60
if i bet $10 35-65
if i bet $25 30-70
if i bet $100 15-85
if i bet $500 5-95
and as bets rise the number of hands the dealer wins in a row increases
you guys are obviously playing other games and have it more closely monitored or keep your bets small to avoid the change in odds.
maybe its just me
but thats what ive found and why i dont play online blackjack any more
i still have to win a hand
if i bet $1 it seems i have 48-52 chance of winning vs losing a single hand
if i bet $5 40-60
if i bet $10 35-65
if i bet $25 30-70
if i bet $100 15-85
if i bet $500 5-95
and as bets rise the number of hands the dealer wins in a row increases
you guys are obviously playing other games and have it more closely monitored or keep your bets small to avoid the change in odds.
maybe its just me
but thats what ive found and why i dont play online blackjack any more
I simply don't trust the online software. With casinos offering 50-200% bonuses, the score should be 27-0. Instead it is 17-10. I trust that most RX posters are not idiots and can play simple blackjack - hell with a 100% bonus, you should be able to come out ahead if you simply dont hit any hands (Or it would be close).
I think 17-10 tells me all I need to know about the software and casinos out there.
-Sean
I think 17-10 tells me all I need to know about the software and casinos out there.
-Sean
Obviously, if the odds are changing with the bet size, the casino is rigged and there's nothing to talk about.
I've never tried to maximize the bet size before. Until reading this thread,
I thought the opposite of what I concluded, that there was no difference between betting large and grindind out a bonus. Obviously I now realize that I was wrong.
So thanks, WVU for opening my eyes. You definitely had the right idea, even though I believe it was for the wrong reason.
I've never tried to maximize the bet size before. Until reading this thread,
I thought the opposite of what I concluded, that there was no difference between betting large and grindind out a bonus. Obviously I now realize that I was wrong.
So thanks, WVU for opening my eyes. You definitely had the right idea, even though I believe it was for the wrong reason.
levistep said:
A group of us paid for the development of this several years ago (big, big bucks). It is very complex as far as card recognition. It was never meant for public use. Just as useful as the bot itself is the database of cards dealt and the statistics that it analyzes. Basically we know who cheats and who doesn't. I need to shut up now.
You said the reason betting bigger has a higher EV because you are betting less against a game with negative expectation, since you are busting out sooner.
I don't believe that this is correct at all, but rather simply because you are using the bonus as a part of your wagering amount, rather than only you're own money, giving you the higher EV.
That might be what you were trying to say, but it didn't sound like it.
I don't believe that this is correct at all, but rather simply because you are using the bonus as a part of your wagering amount, rather than only you're own money, giving you the higher EV.
That might be what you were trying to say, but it didn't sound like it.
WVU, I may have been on the wrong side of variance - I can believe that.
I have more of a problem that my crash always seemed to come right before I hit rollover. I too could get over this.
But you must agree with 100% bonuses, the score should not be 17-10 if the casinos are fair. Maybe they are, but I have my doubts.
Of more importance, WVU, do you have access to people who could write basic, but profitable poker bots? I think this might be the most lucrative thing out there now.
-Sean
I have more of a problem that my crash always seemed to come right before I hit rollover. I too could get over this.
But you must agree with 100% bonuses, the score should not be 17-10 if the casinos are fair. Maybe they are, but I have my doubts.
Of more importance, WVU, do you have access to people who could write basic, but profitable poker bots? I think this might be the most lucrative thing out there now.
-Sean
I've done 8 sites so far and I finished up money (not including the bonus) on every single site except bet365 (busted their because I was betting 3x my normal bet/hand).sean1 said:
That's a lot of hands. Millions of hands back during the glory days of 2x bonus wager requirement must have been so much money. You must have made $300k+.WVU said:
The screen scraping technology is the hardest part with poker bots, the actual coding of the logic is relatively easy. I did a poker bot in Java as part of my final university project (acting as the AI for a J2ME Poker game for mobile telephones), but wouldn't have a clue how to go about screen scraping the clients.
