Blackjack online.... Random or rigged

[gobucs]

I'm a U.S. player. I've played Blackjack on several websites, and I've seen the same disturbing patterns on every single website. I've played PartyPoker (when it was legal in the US), Bodog, and Villento. There were some others, but I don't recall all of their names. These are the main three. The patterns that I'm referring to are numerous, but most notable is the fact that when I bet $1 or $5 a hand I win a reasonable number of those hands. But when I increase the bet to $100 or $200 a hand, I lose an unreasonable number of hands. I've literally played millions of hands and lost well over $100,000 playing Blackjack online. I'm convinced that I'm not getting a fair game. I see delays in the turn cards when a larger bet is made. Almost like the software is searching for a particular card to turn. It sounds crazy I know. But to sit through one session and see what I see on a reoccurring basis leaves me convinced that I'm not getting a fair game. I love Blackjack, it's the only game I care to play, but I don't think I'm ever going to find a fair game online. How can the same patterns exist on every site I've played on? Has anyone else had these issues? Can anyone give me the name of an online casino that accepts US players, and offers a fair game?
GoBucs

[ugaboga]

Have you tried live dealer blackjack?

[PSB]

he patterns that I'm referring to are numerous, but most notable is the fact that when I bet $1 or $5 a hand I win a reasonable number of those hands. But when I increase the bet to $100 or $200 a hand, I lose an unreasonable number of hands. I've literally played millions of hands and lost well over $100,000 playing Blackjack online. I'm convinced that I'm not getting a fair game.

gobucs, how can you verify that you've actually played millions of hands?

I have seen the pattern you mention with small samples, but have also received emails from casino support defending their RNG. Their explanations seem to make sense. Still there are plenty of players who join as new players, deposit $50 to start and walk away with $1500 cashouts. Yet I wonder the overall results of high stakes players who actually recorded all their results in a database/spread sheet of sorts.
Can you elaborate more on that front?
It also seems that $1-5 bets are more easily won in free mode...but increase to $100+ and suddenly the dealer "coincidentally" beats your KK with blackjack.

[SlotsWizard]

I understand your frustration, but I really think most of it boils down to perception.

I cannot count how many times I thought to myself "Self, that felt rigged!" but then I go back and analyze the game logs only to find that the results are surprisingly in line with expectation. I remember at a Rival casino after a streak of seemingly bad luck, I went and checked the logs... the total return was 99.92%, well in line with expectations. Results with Microgaming casinos are similar.

I would recommend restricting your play to casinos that offer you access to your game logs from within the software, without having to jump through hoops of fire to get them. This includes Rival and Microgaming. I'm not sure about others.

Also, call it quits while you're ahead. After a good run, don't start thinking that you can bet bigger and bigger and never experience an equal and opposite downturn in your luck... because you will, eventually, and that's got nothing to do with being rigged and everything to do with the combination of variance and the house edge.

When all else fails and you need some humor, read my New Blackjack Rules.

[lojo]

As well if you chart your play, not by return percentage, but by hands won, lost, or tied - and your 'big bet' sample is large enough - you should see very similar win/lose/push ratio.

But if you chart 10,000 $1,2,5 hands and only 50-100 $100-$300 hands, there can be a large skew, because the big bet sample will not be big enough to truly compare.

[Eliot Jacobson]

Greetings,

When I audit blackjack at on-line casinos, I look for wins/losses/pushes, and also fairness of first card, second card, draw cards, number of hands of each value for the player and dealer, and a host of other metrics (unmentioned to protect the security of the audits).

The monthly audit logs I see may have 500,000 to 1,000,000 hands at some casinos. Even then, I will see RTP's (return to player) between 95% and 102% with a lot of variation in between.

If the THEO (theoretical return to the player) of a blackjack game is 99.5%, I would not expect a reasonable convergence to that number with a simulation of fewer than 100 million hands. Most software programs that model blackjack to obtain numbers accurate to *two* decimal digits, will run for 1 billion hands (or more). For example, see www.qfit.com . The long run is *REALLY LONG* for blackjack.

Based on my personal experience auditing online blackjack, for both casino clients and for players who doubt their fairness, in every case the game offered has been fair. Still, I remain vigilant in my audits. Likewise, you should be vigilant and cautious in your play, but also take into account the incredible variance that happens in these games.

--Eliot

[aka23]

I'm a U.S. player. I've played Blackjack on several websites, and I've seen the same disturbing patterns on every single website. I've played PartyPoker...

In the case of Party Casino, the fairness has been questioned by players quite a few times. In the warning on my site, I list the following dealer upcard stats. The chance of this distribution occurring randomly is quite low. I've also noticed the issue that you mentioned with bet raises, although I didn't attempt enough raises to confirm that it was more than random chance.

Dealer 2 -- 234
Dealer 3 -- 177
Dealer 4 -- 128
Dealer 5 -- 323
Dealer 6 -- 178
Dealer 7 -- 261
Dealer 8 -- 323
Dealer 9 -- 321
Dealer T -- 355
Dealer J -- 291
Dealer Q -- 332
Dealer K -- 305
Dealer A -- 285

Having said that, many players do well with this software (former iGlobalMedia) and have results near expectations. Bodog and Villento use different software, and I am not aware of similar issues with their softwares.

But when I increase the bet to $100 or $200 a hand, I lose an unreasonable number of hands.

Via personal play, I have an excellent history with large bets (as much as £400 per hand) on Microgaming software. I am definitely on the positive side of variance.

[Eliot Jacobson]

In the case of Party Casino ...

Dealer 2 -- 234
Dealer 3 -- 177
Dealer 4 -- 128
Dealer 5 -- 323
Dealer 6 -- 178
Dealer 7 -- 261
Dealer 8 -- 323
Dealer 9 -- 321
Dealer T -- 355
Dealer J -- 291
Dealer Q -- 332
Dealer K -- 305
Dealer A -- 285

If these numbers are accurate, the odds are roughly 1 in 10000000000000000000000000000000000000000 that this (or a worse result) will occur purely by chance. (1.2094E-40).

I have no way of verifying your numbers, I am just telling you the results of a simple chi-squared test.

--Eliot

[Eliot Jacobson]

Based on the post by aka23, I decided to log on to my Party Gaming account and play blackjack to investigate "first card" bias.

I played their Blackjack - Single Player (Las Vegas Downtown Blackjack) game, in practice mode, flat betting $1 per hand. My play started at 6:05 PM and lasted until 7:20 PM. During this time I played 489 hands and recorded the dealer up-card for each of these hands. These results were:

2 -- 32
3 -- 41
4 -- 36
5 -- 31
6 -- 30
7 -- 36
8 -- 41
9 -- 38
T -- 50
J -- 40
Q -- 38
K -- 36
A -- 40

These results are statistically normal. The expected number for each card was 37.62. The chi-squared statistic is 8.75, which means these results are effectively random. I find no "first card" bias in this game, as I played it, at Party Gaming casino.

--Eliot

[aka23]

The numbers I quoted are based on data posted on another forum. He played 3513 hands of Party multi-player blackjack in June 2007, flat betting $1 per hand. He lost ~$1000 over the course of this wagering, then analyzed the game logs and posted the upcard data. He probably still has the game logs, so he may be able to send them to you. I'll PM you contact data.