Video Poker stats

I was told that online video poker is played off software that is based on random 52/53 card deck draw. If this is true then in time a player should expect with perfect play expected probabilities.

I decided to give a test at a reputable RTG online casino. If in fact video poker is not being played from software based on pre-determined house edge advantage like slot machines, then in fun play it should be the same type of results as in real play.

I played only Joker Poker and started each session regardless if fun play or real play with a $500.00 bankroll, and each hand was played at $5.00 per hand. I play approx. 19 to 23 hands per minute. Usual session would last between 15 to 30 minutes before fatigue sets in or bankroll is zero, whichever came first.

Before I give you the total sessions and hands I just want to mention that after my first 8 sessions in fun play I had exactly 0 losses. My bankroll was exactly at the original starting point. After 8 sessions of playing for real money I was losing -$3,000.00 dollars.

Here are the total numbers and could you please tell me what the RTP would be for each mode?

FUN PLAY:

15 sessions, 7,117 hands, negative -$1,000.00 dollars

REAL PLAY

08 sessions, 3,325 hands, negative -$3.000.00 dollars

In addition all hands combined only three Joker Royals were hit and two 5oak were hit. One of the 5oak’s was hit in real play and the other 4 premier hands were hit in fun play.


Thank you

the number of trials you are listing is fairly insignificant.

online casinos have been caught cheating in live poker. I generally don't trust them and don't play online.

The general question of online casino cheating HERE

We'd have to know what the paytable was, to get your expected value. It would also be helpful to know if you were playing the proper strategy.

In your "real play" experience, you bet $16,625, and lost $3000. A bad result, but not off the charts by any means. A bad paytable+poor strategy could easily account for half of that loss, with variance accounting for the other half.

Quote: mkl654321

We'd have to know what the paytable was, to get your expected value. It would also be helpful to know if you were playing the proper strategy.

In your "real play" experience, you bet $16,625, and lost $3000. A bad result, but not off the charts by any means. A bad paytable+poor strategy could easily account for half of that loss, with variance accounting for the other half.

Seq. Royal - 10000
Royal - 4000
5OAK - 750
Royal W/Jo - 400
Str. Flush - 250
4OAK - 100
Full House - 35
Flush - 25
Str. - 15
3OAK - 10
Two Pair - 5
K's or better - 5

I play 100% perfect strategy unless mis-click which is very rare. With total hands above maybe 3 mis-clicks throughout.

Quote: 4ofaKind

Seq. Royal - 10000
Royal - 4000
5OAK - 750
Royal W/Jo - 400
Str. Flush - 250
4OAK - 100
Full House - 35
Flush - 25
Str. - 15
3OAK - 10
Two Pair 5
K's or better 5

I play 100% perfect strategy unless mis-click which is very rare.

This paytable should return about 100.4% (this is a rough calculation; the fullpay casino paytable is the same as this one, except 5OAK pays 1000 and Joker Royal pays 500, so I'm estimating the effect of those hits--the fullpay casino game pays 100.65%). I don't add anything for the sequential royal, as it's an extremely rare hand and the bonus is only 6000 coins anyway.

Quote: mkl654321

This paytable should return about 100.4% (this is a rough calculation; the fullpay casino paytable is the same as this one, except 5OAK pays 1000 and Joker Royal pays 500, so I'm estimating the effect of those hits--the fullpay casino game pays 100.65%). I don't add anything for the sequential royal, as it's an extremely rare hand and the bonus is only 6000 coins anyway.

Could you tell me what the RTP was with each play mode separately?

Quote: 4ofaKind

Could you tell me what the RTP was with each play mode separately?

RTP?

Quote: mkl654321

RTP?

Return to player percentage. Not sure if that's the right term for video poker.

Quote: 4ofaKind

Return to player percentage. Not sure if that's the right term for video poker.

No, it isn't. The proper term is "expected value", or "EV", expressed either as an absolute percentage (100.4%), or a difference from 100% (+0.4%).

If the paytable is the same, the expected value is the same for either "fun play" or "real play".

Quote: mkl654321

... If the paytable is the same, the expected value is the same for either "fun play" or "real play".

mkl, you are assuming that the casino is offering an honest game. I think that the original post in this thread was an indication that 4ofaKind was suspicious that the online casino might indeed not be playing it quite right; i.e., might not be playing with a full deck, ignoring the common implication of that phrase.

Quote: mkl654321

No, it isn't. The proper term is "expected value", or "EV", expressed either as an absolute percentage (100.4%), or a difference from 100% (+0.4%).

If the paytable is the same, the expected value is the same for either "fun play" or "real play".

So then at this point of the samples above what was the difference from EV ? Hope I'm asking the right question here.

Quote: Doc

mkl, you are assuming that the casino is offering an honest game. I think that the original post in this thread was an indication that 4ofaKind was suspicious that the online casino might indeed not be playing it quite right; i.e., might not be playing with a full deck, ignoring the common implication of that phrase.

Well, I decided to answer the question "straight up", because if the online casino is cheating, the effective return is 0%. But the set of hands played was far too small to draw such a conclusion, nor would the reported results be all that unusual in a completely honest game. So the results reported are not nearly enough to imply dishonesty on the part of the online casino. I wouldn't send them five cents to play with, because I think you'd have to be crazy to trust an online casino, but the lousy results the OP had are not ipso facto proof of cheating.

mkl, I agree with you. Both about the lack of sufficient data and about the unattractiveness of online gaming. Wait a sec, though -- are we dissing the whole industry of the only paid advertiser for "our" forum? Shame on us.

Quote: Doc

mkl, I agree with you. Both about the lack of sufficient data and about the unattractiveness of online gaming. Wait a sec, though -- are we dissing the whole industry of the only paid advertiser for "our" forum? Shame on us.

I think that those people who are reluctant to send actual money to some anonymous gambling site, located in an unknown location, and play casino games with that money, the honesty of which cannot be verified, are in the minority, even here. Or to put it another way, I don't expect anyone to slap their forehead and say, "Great Scott! Online gambling is an extremely foolhardy pastime!" because of anything you or I say.

I know this is a small sample to conclude anything with fact, but with my limited math abilities I have fun play results returning 98% and real play returning 82% at this point. I find this unusual since there should be no optional settings for video poker in play mode if in fact software is based on 52/53 card deck draws. (not pre-determined house edge advantage settings like slots) I'm prepared to risk another 2K in four more real play sessions, then see how long it takes fun play to equal the same. I'm having fun doing this and am willing to write off the 5k losses off as an experiment.

I guess the amount needed that would make most people even remotely involved with online gaming satisfied with sufficient test results to come to any serious conclusion; I'm not prepared to loss that much trying to prove a point.

Quote: 4ofaKind

I know this is a small sample to conclude anything with fact, but with my limited math abilities I have fun play results returning 98% and real play returning 82% at this point. I find this unusual since there should be no optional settings for video poker in play mode if in fact software is based on 52/53 card deck draws. (not pre-determined house edge advantage settings like slots) I'm prepared to risk another 2K in four more real play sessions, then see how long it takes fun play to equal the same. I'm having fun doing this and am willing to write off the 5k losses off as an experiment.

I guess the amount needed that would make most people even remotely involved with online gaming satisfied with sufficient test results to come to any serious conclusion; I'm not prepared to loss that much trying to prove a point.

Even what you are doing seems sort of like devouring a bucket of freshly-picked mushrooms to see if they're poisonous.

I would suggest that you do the following on your next "real play" session: track the frequency of the Joker when dealt/drawn. On the initial deal, the Joker should be dealt a little less than one time in ten (5/53). On the draw, it should be dealt once out of every 48 cards drawn. One of the most obvious ways for an online Joker VP game to cheat would be to decrease the frequency of the Joker. This is also a greater number of trials than the number of hands you play, so the data would be more significant than your actual play results.

Fun game: -1000 of 35385 bet or .028 or 2.8%
Real game: -3000 of 16625 or .180 or 18%.

Apply the standard statistical z-test for two proportions with null hypothesis that the two games are equal in loss proportion against the alternative hypothesis that the two games are not equal in loss proportion.

The test statistic is z=27.24 which is highly significant.

Conclusion: the two games are not statistically the same at <.01 level.
--------------------------------------------------------------------------------

test statistic: z = |p1-p2 |/s

where:

p1 = proportion 1
p2 = proportion 2

s = sqrt(p(1-p)/n1 + p(1-p)/n2) p = (p1n1+p2n2)/(n1+n2) n1 = sample size 1 n2 = sample size 2

significance test:
z >
2.576 for 99% level of confidence
1.96 for 95% level of confidence
1.645 for 90% level of confidence


Matilda

Thanks Matilda for that information. Actually the only thing I could clearly understand was the Conclusion: "the two games are not statistically the same." Can you tell me if my percentage of return was in order?

I also will now keep track of Joker pulls and report again when I complete the test.

Quote: mkl654321

This paytable should return about 100.4% (this is a rough calculation; the fullpay casino paytable is the same as this one, except 5OAK pays 1000 and Joker Royal pays 500, so I'm estimating the effect of those hits--the fullpay casino game pays 100.65%). I don't add anything for the sequential royal, as it's an extremely rare hand and the bonus is only 6000 coins anyway.

Wrong guess again, and there's so much wrong with it that I'm embarrassed to point it all out. Here's the Kings-or-Better paytable IN FULL PAY FORMAT from the vpgenius site. As you can see, even with a 50,000 credit Sequential RF where the bonus you claim as an expert to be "only 6000 coins" is actually 46,000 coins (which I've seen in casinos) the game expectation is 100.1815%.

Sequential Natural Royal Flush $50,000.00 81,266,463,720 0.0000% 1 in 4,353,476.7887 22.9655 0.2297%
Natural Royal Flush $4,000.00 8,249,695,147,800 0.0023% 1 in 42,885.4227 14.8861 1.8654%
Five of a Kind $750.00 33,022,151,426,880 0.0093% 1 in 10,713.7678 2.0721 1.4001%
Wild Royal Flush $400.00 35,283,314,276,640 0.0100% 1 in 10,027.1664 0.6224 0.7978%
Straight Flush $250.00 203,144,703,539,040 0.0574% 1 in 1,741.5746 1.3785 2.8710%
Four of a Kind $100.00 3,024,712,518,379,200 0.8549% 1 in 116.9670 3.0858 17.0988%
Full House $35.00 5,545,119,257,821,440 1.5673% 1 in 63.8024 0.5639 10.9714%
Flush $25.00 5,545,241,720,755,200 1.5674% 1 in 63.8009 0.2506 7.8369%
Straight $15.00 5,890,396,434,131,520 1.6649% 1 in 60.0625 0.0665 4.9948%
Three of a Kind $10.00 47,367,334,189,517,760 13.3885% 1 in 7.4691 0.1334 26.7770%
Two Pair $5.00 39,252,560,873,137,920 11.0948% 1 in 9.0132 0.0000 11.0948%
Kings or Better $5.00 50,393,338,901,125,920 14.2438% 1 in 7.0206 0.0000 14.2438%
Nothing $0.00 196,493,178,479,556,960 55.5392% 1 in 1.8005 0.5574 0.0000%
Totals 353,791,663,505,280,000 Any Win: 1 in 2.2492 46.5822 100.1815%