MG Blackjack tweeking re-visited

Arismac

Dormant account
Joined
Nov 22, 2015
Location
Mandurah, Western Australia
I have just finished playing 100 hands of MG "Multi-hand Vegas Strip" Blackjack. I had three upstanding and honest friends with me. We divided the hands up between two Casino's belonging to two different groups.

The Casino "busted" (went over 21) on just 9 hands. On 91 hands the Casino turned hands of between 17 and 21. We calculated that this gave the Casino's a winning edge of greater than 5%. We tried with both Practice Money and Real Money and got very similar results. I really believe there is more than "Lady Luck" messing with the game.

I think it is time some bright person produced a very different Blackjack Strategy card for players at some online casino's.

Cheers, Mac
 
I perhaps should have mentioned that I took 100 screen grabs this time just in case there are any doubters.

I guess what concerns me is if this goes on with one game how many more games get "tweaked" to suit the casino? Perhaps I should send the screen grabs to eCOGRA?
 
I perhaps should have mentioned that I took 100 screen grabs this time just in case there are any doubters.

?

It is too small a sample, your result is well within statistical expectations IMO.

I once had a session where my return was 82 % over 600 hands. That's roughly 17.5 % for the dealer, your 5 % is marvelous compared to that :D.

On a serious note, you cannot expect the 99.5 % return in the game of BJ on a small sample.
Because the house edge of ca 0.5 % (or even less in some versions) includes, for example, situations where you split and double and win four times your stake. And these situations are quite rare. So it can easily happen that these situations favourable for a player do not occur over an extended session, which results in a much lower return to player than 99.5 %.
 
I spent a while looking into the MG Blackjack.

One experiment had me set the game to autoplay using the MG provided default strategy and playing almost 100,000 hands at £1. I took readings every 10,000 hands, and to my amazement the experiment ended with me being UP by an astonishing £700+. This meant a swing of over 700 bet units in the WRONG direction over quite a decent sample, although still a small sample in statistical terms. I thought something had gone wrong and it was a bug, so I decided to up the stake to £10 and see if the trend would continue over the next 100,000 hands. Of course, it didn't, and the first run wiped out my £700.

The beauty of MG Blackjack is that the strategy tables can be edited by the player so that they can autoplay to their preferred strategy if they feel the game doesn't behave as it should. In the "good old days", MG Blackjack on autoplay was widely used to beat boni.

The autoplay is now a curiosity, no good for beating a bonus as it's probably banned in the terms, but still useful for "wager challenge" contests, although most are now held on slots.

It might be a fun exercise to analyse a good run of MG Blackjack, and then try out an edited strategy on autoplay to see if the game can be beaten by factoring in the "cheat" that seems to allow the dealer to avoid busting as often as they should.
 
Many thanks for a very interesting read.

As an update I have just finished placing 25 real money bets with two MG powered casinos playing Multi-hand Vega Strip.

Casino #1 - Casino did not "bust" once and of course I lost.
Casino #2 - Casino "busted" twice and of course I lost.

As I would think that very few players would play 100's of hands on end and even less players would play 1,000's hands on end my "rule" now is:-

If the casino does not "bust" more than once out of the first ten deals, walk away and try another brand of software because the odds are stacked against the player.
 
Just for the record we ran another batch of 100 hands to the casino which is 500 player hands in Multi-player.

The Casino busted on 17% of hands and the Player on 24%. A very big edge to the Casino.

A friend who works on the floor at an Australian BM Casino described the result as "smelly".
 
Just for the record we ran another batch of 100 hands to the casino which is 500 player hands in Multi-player.

The Casino busted on 17% of hands and the Player on 24%. A very big edge to the Casino.

A friend who works on the floor at an Australian BM Casino described the result as "smelly".

You might wanna check this guy out. He has BJ analysis for mg casinos.

You do not have permission to view link Log in or register now.
 
The Casino busted on 17% of hands and the Player on 24%. A very big edge to the Casino.

I think that is what is supposed to happen; the higher rate of busting on the part of the player is outweighed by the BJ paying 3:2 and by the split and double options, which the player has and the dealer does not have.
 
A friend who works on the floor at an Australian BM Casino described the result as "smelly".

Believe me, this comment by your friend really does not carry any weight at all.
This involves suprisingly high-level mathematics (at least from my point of view; I do not claim to understand the mathematics but I read some posts over the years here, from which I got this impression); that is why only real experts are able to determine whether there is something fishy about a specific online game of BJ or not.
 
It might be a fun exercise to analyse a good run of MG Blackjack, and then try out an edited strategy on autoplay to see if the game can be beaten by factoring in the "cheat" that seems to allow the dealer to avoid busting as often as they should.

I am not sure if I fully understand what you are saying but I consider writing the words "to see if the game can be beaten" in a post responding to an obviously new player to be very dangerous.

Due to your amount of time in this forum, he could put trust in your comment, which IMHO is really way out of line and really misleading because IMO suggesting that an edited strategy, i.e. a strategy deviating from the mathematically ideal strategy, could produce better results is a complete nonsense and could put the OP on a dangerous path.

In a nutshell, of course the game cannot be beaten. It has a built-in house edge and that is that. Any deviation from ideal strategy increases the house edge. Period.
 
I am not sure if I fully understand what you are saying but I consider writing the words "to see if the game can be beaten" in a post responding to an obviously new player to be very dangerous.

Due to your amount of time in this forum, he could put trust in your comment, which IMHO is really way out of line and really misleading because IMO suggesting that an edited strategy, i.e. a strategy deviating from the mathematically ideal strategy, could produce better results is a complete nonsense and could put the OP on a dangerous path.

In a nutshell, of course the game cannot be beaten. It has a built-in house edge and that is that. Any deviation from ideal strategy increases the house edge. Period.

A fair point, but Microgaming are the ones who put this ability into the Blackjack autoplay system, and clearly they thought that players would like to do such things.

However, there is always the option to run such experiments in "fun mode" rather than commit real money, and I have also carried out many of these experiments in "fun mode" so that I am free to push the boundaries whilst not actually risking my own money.

Whilst you can't beat a properly functioning game with a mathematical house edge simply by changing strategy, there has been some doubt about the default Blackjack strategy tables being the best, and Microgaming have updated some from their original settings as it turned out they were not optimal.

If you DO appear to beat Blackjack using an edited strategy table, then the main 2 possibilities are that the sample size is too small, or the game is malfunctioning. Online casino games DO malfunction, and occasionally this is in favour of the player.

You don't necessarily need to get the RTP over 100% in order to beat the house long term, one has to factor in whether the play generates loyalty points, or can contribute value through play in conjunction with a wagering promotion run by the casino. If the sum total of the value from all these different things is greater than the equivalent of an RTP of 100%, then you CAN beat the casino, until of course they spot the error in their promotional setup and put a stop to it, which normally happens once too many players cotton on.
 
Like vinyl and some other members said 100 hands is way too small of a sample. 100 thousand hands, and there should be nearly correct results, albeit even that is a bit small sample. The game like all games including slots can swing massively to one direction, only to come back much later on to the normal RTP.
 
A fair point, but Microgaming are the ones who put this ability into the Blackjack autoplay system, and clearly they thought that players would like to do such things.

However, there is always the option to run such experiments in "fun mode" rather than commit real money, and I have also carried out many of these experiments in "fun mode" so that I am free to push the boundaries whilst not actually risking my own money.

Whilst you can't beat a properly functioning game with a mathematical house edge simply by changing strategy, there has been some doubt about the default Blackjack strategy tables being the best, and Microgaming have updated some from their original settings as it turned out they were not optimal.

If you DO appear to beat Blackjack using an edited strategy table, then the main 2 possibilities are that the sample size is too small, or the game is malfunctioning. Online casino games DO malfunction, and occasionally this is in favour of the player.

You don't necessarily need to get the RTP over 100% in order to beat the house long term, one has to factor in whether the play generates loyalty points, or can contribute value through play in conjunction with a wagering promotion run by the casino. If the sum total of the value from all these different things is greater than the equivalent of an RTP of 100%, then you CAN beat the casino, until of course they spot the error in their promotional setup and put a stop to it, which normally happens once too many players cotton on.

"Beating the house" and "beating the game" are completely different things.

I apologise if I was too harsh on you but given your status in this forum and your wealth of experience with gambling, I thought you should know better than to post things like that in a response to a newbie.

Otherwise, I appreciate your other points you make in your post, save for one, which is a little inaccurate IMO: If one seems to be "beating the game", the ONLY 2 possibilities are the variance (sample being too small) or the malfunctioning of the game.
 
Interesting debate this one and to throw further thought into this, I used to play Blackjack a few years back at Ladbrokes. Not day in day out, just a few days every 6 months to stop any addiction issues, etc. My experience as follows:

Initially just played for the sake of it, albeit quite cautious (sticking on 12's and 13's in fear of busting, etc), then tried a strategy. My theory was that laws of averages would provide a winner after say 3 or 4 losing hands, so bet £2 a hand and played to lose, then a hand of say £20 which many times actually won. First few times of doing this I made a fair bit of money, then upped the stakes to £5 losing hands then anything up to max bets of £200 (far out of my comfort zone, so not done very often). More often than not I won on the higher stake (got blackjack a couple of times yielding £500 win) and one time actually stuck on 12 with dealer turning over a 10, then 6 then busted out with another 10 while on a £200 stake. Seriously, I was a bag of nerves doing £200 stakes!!

After a couple of years playing with breaks of 6 months between visits, I was up over £4k from quite low deposits (less than £200 in total).

Ok, not loads of games to base any predictions on and probably coincidence and getting lucky, but at the time it did work by losing a few hands on purpose (within reason), upping stake and then play to win.

Would love to get my game history up to prove this, but fear it is too far in the past now, plus am locked out of Ladbrokes (had two different accounts - one with debit card registered on and the other with a credit card - went some time playing like this then accounts got locked).

Being older and wiser now, I probably wouldn't try the same again in fear of forfeiting any winnings due to 'irregular play' !!!
 
After another day of very entertaining research we are starting to create a "barometer" of whether the MG software is going to be nice or nasty. It seems as though there is a correlation between percentage of times the Casino busts and if the Player is likely to win or lose.

Without going into figures yet because this is still very much a work in progress, it seems obvious that the higher percentage of Casino busts the better the return to player. What we can not yet work out is why the percentage will quite suddenly vary by as much as 10%.

The "sudden change" appears common to the several MG powered casinos we have been researching.

More to come.
 
After another day of very entertaining research we are starting to create a "barometer" of whether the MG software is going to be nice or nasty. It seems as though there is a correlation between percentage of times the Casino busts and if the Player is likely to win or lose.

Without going into figures yet because this is still very much a work in progress, it seems obvious that the higher percentage of Casino busts the better the return to player. What we can not yet work out is why the percentage will quite suddenly vary by as much as 10%.

The "sudden change" appears common to the several MG powered casinos we have been researching.

More to come.

You should do much of the work in the "fun mode" setting as you can freely run experiments without having to risk your own money.

When I ran a Blackjack trial for real money I set the stake to £1, the lowest possible. After my first run produced a profit of over £700 at the £1 stake, I tried to repeat the experiment. The second run did not produce the amazing experience of going £700 up, in fact it busted out my deposit.

During the run that ended with the £700 profit, there were swings of over £100 in either direction, and this would explain why I lost the £700 when trying a run at £10 stake, I only had 70 betting units available, so one swing of -100 would have wiped out my bankroll, and that's what happened.

My initial thought that there might be a bug in the game that would have allowed me to carry on winning over 200,000 or 300,000 hands to a similar level were brought down to Earth in the second attempt at a £1 stake where I wasn't able to produce the same result despite selecting the same game, same stake, and same casino.

It just shows that you CAN have a remarkably positive result over what you would consider a very large sample of 100,000 hands, only to find that it's still such a small sample in terms of the game's mathematics that the astounding result was merely one of the luckier spells in the even longer set of samples it would take for the game to reveal it's true TRTP, which for Blackjack is in the order of millions of hands.

One aspect that makes online Blackjack behave differently to the game when played in a live casino is the lack of shoe penetration in the computer version. Each hand will start with what is effectively a freshly shuffled deck, so seeing all the ten cards used up in the first few hands doesn't mean there is less chance of ten cards cropping up in later hands.
 
Thank you for yet more very interesting contributions.

I initially set out to see if "Practice Mode" and "Real Money" would produce similar results. We had two players playing with AUD and two playing "Fun Money". Over 100 games each results were close enough to call them "same", so that satisfied us that we can experiment without spending too much of the kids inheritance.

Results on a day to day basis have been very different however. I have just finished playing 65 games and "busted" 81 times to the casino's 65. As a result I lost A$100.00. However at 50 games player 52 and casino 50 and I was exactly square, neither winning or losing.

While I accept arguments regarding 1,000's or even millions of hands these figures have no practical value for the "mug punter" which is the class I fit into. What I am trying to do, with your help, is attempting to gauge the "mood" of the software.

If it could be left exactly as it was half an hour ago I would rate it to be a "Fair Bet" but this is certainly not always the case.
 
Thank you for yet more very interesting contributions.

I initially set out to see if "Practice Mode" and "Real Money" would produce similar results. We had two players playing with AUD and two playing "Fun Money". Over 100 games each results were close enough to call them "same", so that satisfied us that we can experiment without spending too much of the kids inheritance.

Results on a day to day basis have been very different however. I have just finished playing 65 games and "busted" 81 times to the casino's 65. As a result I lost A$100.00. However at 50 games player 52 and casino 50 and I was exactly square, neither winning or losing.

While I accept arguments regarding 1,000's or even millions of hands these figures have no practical value for the "mug punter" which is the class I fit into. What I am trying to do, with your help, is attempting to gauge the "mood" of the software.

If it could be left exactly as it was half an hour ago I would rate it to be a "Fair Bet" but this is certainly not always the case.

Or college funds either!

A mug punter wouldn't care to doing the analysis your doing.
 
The trend remains the same. Casino busted 110 times Player busted 134 times. Player lost 14.7% of opening bank.

In order to make this realistic it is possibly important to know the Strategy used.

Casino and Player sit on 17 or better
Player must hit 16 or less
No splits
No doubling
No insurance
No surrender
Re-deal any hand with Blackjack

Level stake. (This is not how I normally bet)

So I consider that using a Strategy Card would move the edge closer to a loss of about 5% in this example.

It is starting to look as though a v$ Practice of 15 - 25 hands may be enough to provide a clue as to what mood the Casino is in. Obviously the longer the trial the accurate the result.

Perhaps there will be a stronger indication after a few 100 more hands. So if a trial, using v$, returned this result I would recommend perhaps doing some gardening and coming back later.
 
The trend remains the same. Casino busted 110 times Player busted 134 times. Player lost 14.7% of opening bank.

In order to make this realistic it is possibly important to know the Strategy used.

Casino and Player sit on 17 or better
Player must hit 16 or less
No splits
No doubling
No insurance
No surrender
Re-deal any hand with Blackjack

Level stake. (This is not how I normally bet)

So I consider that using a Strategy Card would move the edge closer to a loss of about 5% in this example.

It is starting to look as though a v$ Practice of 15 - 25 hands may be enough to provide a clue as to what mood the Casino is in. Obviously the longer the trial the accurate the result.

Perhaps there will be a stronger indication after a few 100 more hands. So if a trial, using v$, returned this result I would recommend perhaps doing some gardening and coming back later.

These are the worst rules I've ever seen!
 
These are the worst rules I've ever seen!

I think the OP was just playing like this for the sake of comparison of the number of busts on dealer's and player's side.

But all this "testing" is absolutely pointless anyway...

For the OP:
I do not know how to put this tactfully and I mean it in the nicest way but "gauging the mood of the casino" or the "mood of the software" is absolutely delusional. Past results have no bearing on the future results in an online casino.

I would give you more mathematical examples of what blackjack can do, what swings it is capable of but VinylWM already posted something about that and you said you find these arguments irrelevant.
I tried to help you but under these circumstances, I give up, sorry.
 
Thank you for yet more very interesting contributions.

I initially set out to see if "Practice Mode" and "Real Money" would produce similar results. We had two players playing with AUD and two playing "Fun Money". Over 100 games each results were close enough to call them "same", so that satisfied us that we can experiment without spending too much of the kids inheritance.

Results on a day to day basis have been very different however. I have just finished playing 65 games and "busted" 81 times to the casino's 65. As a result I lost A$100.00. However at 50 games player 52 and casino 50 and I was exactly square, neither winning or losing.

While I accept arguments regarding 1,000's or even millions of hands these figures have no practical value for the "mug punter" which is the class I fit into. What I am trying to do, with your help, is attempting to gauge the "mood" of the software.

If it could be left exactly as it was half an hour ago I would rate it to be a "Fair Bet" but this is certainly not always the case.

Which is EXACTLY what we would expect from a random game of Blackjack.

The idea that one should leave the casino if the software is in a "bad mood" is more about the human side of gambling. It creates a strategy that pushes the player into taking a break and coming back later rather than getting sucked into the "chase of losses" that is the downfall of many.

Upon return, the software may well be "in a better mood", but this can be explained by the random nature of the game, individual sessions can have very different characteristics, and it's all down to the maths of a random system, rather than the software having pre programmed "moods".

If the software DID have a non random component, a smart player would clean out the casino, so implementing non random games would be VERY dangerous for a casino. If you could predict the "mood" of Blackjack from an initial 15 hands, you would be able to clean out the casino over the long term.

Someone has tried this before, the notorious "Cipher" and his Blackjack predicting and beating software. All those who got sucked into his scheme ended up getting badly burned. He began to ask people to just give him money and he would "invest it" in his team of in-house players who would play Blackjack on their behalf and the winnings, less commission for Cipher, would be returned. Funnily enough, it was at this point that his sure fire system failed to deliver, and all the money was lost at Blackjack (so he claimed at the time).

His software and system APPEARED to work over a short "mug punter" session of relatively few hands, but in reality it didn't, it was all smoke and mirrors and clever marketing.

Treat it as a bit of fun, but don't expect to find the "Holy Grail" at the end of your research.
 
Results on a day to day basis have been very different however. I have just finished playing 65 games and "busted" 81 times to the casino's 65. As a result I lost A$100.00. However at 50 games player 52 and casino 50 and I was exactly square, neither winning or losing.

Given the sample sizes, it sounds like you may be falling into a cognitive bias class called apophenia (the human tendency to see patterns in random data - the foundation of the gambler's fallacy). The only way a game has "mood" is if it is compensated (as in UK AWP fruit machines) or is random but defective.

I think vinylweatherman and Janek12 have summed it up :)
 

Users who are viewing this thread

Meister Ratings

Back
Top