+ Register

[Drazen]

I have 53% strike rate on average odds of 2.03 so it seems I have an slight edge anyway. But I use mild progression to boost my profit.

[WorldBaccaratKing]

is this gonna be another thread like the one with 30 pages with useless mumbo jumbo? or is someone gonna start posting some ideas for reference and some proof about these so-called claims....?

otherwise, its the same story different day. one guy now says he makes 6x his salary but got a job? did i read that correctly?

[WorldBaccaratKing]

i just got new boots in the mail friday, guess i better put them on in case stuff starts getting deep in here.................

[AsymBacGuy]

Nice post (starting from the post's name  , nice replies.

I've been always taught that no any MM ever invented could beat itlr a 50/50 game, let alone a 50/50 taxed game.
Maybe someone have succeeded in that, maybe he/she should present his/her work at the MIT or at The University of Nevada, Reno NV where stuff like this is taken in deep consideration.

IMO and in the opinion of my data, if baccarat would be the product of a perfect 50/50 game (no third card rule) there's almost no way to get an edge. I sayed "almost" because in some situations the study of certain final patterns might produce interesting betting circumstances.

Hence IMO and in the opinion of our data we should focus our attention on the flaws (disparities)the game is constantly although whimsically providing.

The whimsical aspect, IMO and in the opinion of our data, is the hardest part to control. Obviously.

The whimsical aspect constantly fights with the expected values, but itlr (and fortunately even on shortest denominations) the latter will prevail over the former.

In such disorganized world we should be able to spot the situations where the more likely situations (because at bac there are more likely situations unlike other similar games) are more likely to happen.

Is this task an easy work?

Nope, by any means.

The way is relatively difficult to find profitable black jack opportunities, the like baccarat works.
Variance will steadily act, either in positive fashion or in negative fashion.
The paramount difference between those two games (aside the mathematical aspect) is that positive periods are always profitable for the bj player whereas positive periods at baccarat most of the time are misleading; conversely, negative periods for a bj player are always negative, at baccarat negative periods whether properly assessed might be propethic of some positive fragments' production.     

But we baccarat players want to control the uncontrollable, guessing the unguessable, trying to get a balancement on our W/L ratio within too restricted terms.

Bayes insists on the RTM importance.

I can't disagree. Somewhere in my useless post I cited the "regression effect". Along with other parameters, the regression effect will shift the outcomes in our direction.

Naturally, to accept a possible RTM influence automatically presume to wait favourable circumstances. And they are not so frequent, IMO and in the opinion of our data.

We shouldn't forget that what shifts a 50/50 game into a 50.68/49.32 game is the asymmetricity. And this works by a less than 1/10 average impact. Secondary importance will take the card distribution, itself altering the as production.

Easy to notice that such unbalancing forces cannot regularly act, not per every shoe' section, not per every shoe, not per every dozen of shoes and sometimes not even per every hundred of shoes.

IMO, considering beatable a given game itlr means to get an edge by a flat betting mechanical play where no one expert of the world could disagree on that.

Several years ago I made an experiment putting in a dark bowl 52 winning red balls along with 48 blue losing balls, asking my friends to pick up one ball and registering the outcomes. Everytime the chosen ball was reintroduced just to get the same 52/48 WL ratio.
I asked my friends to set up a play on red balls.
Utilizing a bankroll of 50 units, eventually they all went broke, despite their 4% advantage.

as. 



     



 

   



     

[Bayes]

@ Drazen,

Nice to hear you're doing well at sports betting. I think you'll like my new software, I've been tweaking it for months and it's really awesome (although I say so myself).

@ Tomla,

Bayes on your theory  on RTM do you flat bet or use a progression?

It's hard to say because it would depend on your win target, which can be adjusted to 1 unit per X spins. I generally set X = 7 which is fairly conservative. At this rate, I do need to use a mild progression at times (usually not requiring a stake of more than 5 units). I haven't done any extensive testing on this, but it may be possible if you set X = 12 or higher.

As,

Naturally, to accept a possible RTM influence automatically presume to wait favourable circumstances. And they are not so frequent, IMO and in the opinion of our data.

Correct, and that's largely what my software is for - to find favourable circumstances without have to wait hours. The price you pay for this is that the extreme events can be very extreme (because that's what the software is looking for) and this was a fault of the earlier software I used, but these problems have now been largely ironed out because it now searches for more diverse patterns of losses (the diversification principle).

The significance of RTM is not so much that it predicts winners but  rather that it limits losses and keeps stakes low (which should be the prime directive). The logic is really very simple, although some will deny it because it smacks too much of the gambler's fallacy, although it isn't: Assuming the worst sequence from hell can run to 5 standard deviations (confirmed by me in millions of trials) and that such a sequence can come at any time - which would you rather have to cope with? 5 standard deviations  of losses or at most 2 standard deviations? the latter, of course. Therefore, start your current play at 3+ standard deviations. You do the math.

The typical objection is that being independent trials, it doesn't matter what the deviation is before you start playing, none of what has gone before counts for anything. According to these math dogmatists it's just as likely that you will encounter 5 standard deviations on top of the current 3 sd giving 8 sd, and so on to an infinitely receding horizon of standard deviations like the pot of gold at the end of the rainbow which you never actually get any closer to. I prefer common sense and more importantly, experience.

[horus]

I find it fascinating that two completely different approaches to the game can both be profitable showing that the game is not as impossible as some people like to make out. I have to admit that I have not really looked at RTM much. I am more in the XXVV camp where using short cycles to get in and out can be very profitable if you have several angles to attack.

RTM and sports betting does sound interesting.

Southampton football team would be a good example. They are not one of the elite like Chelsea, Man Utd, Man City or Arsenal. Yet they had a great start to the season winning something like 6 games on the bounce. So a team like Southampton with a 6-0 record are not going to finish the season 44-0  This is one area I was reading up on last night. Punters then tend to overestimate the teams record for the rest of the season. So there could be value in opposing Southampton. They did then go on to have a bad losing spell of around 5-6 games before recovering again.

Horseracing sounds like a good vehicle for RTM as well. A horse may build up a winning sequence of 111 but suppose the horse is now at a weight that it has never won at before. Is it going to win it's next few races as well. Was it a lucky winner in it's last race by maybe a horse falling.

All very interesting and I now have a new approach to get my teeth into. 

[Agesta1]

Hi!
This is for regression to the mean and football Wagers.

http://economics-files.pomona.edu/GarySmith/footballOdds/footballOdds.html.

agesta

[Sputnik]

Correct, and that's largely what my software is for - to find favourable circumstances without have to wait hours.

Would you say that each event has the value of 1 and there is two oppisite events (1 in 2).

Would you say this chart is correct for overrepresented and underrepresented events?

The price you pay for this is that the extreme events can be very extreme (because that's what the software is looking for) and this was a fault of the earlier software I used, but these problems have now been largely ironed out because it now searches for more diverse patterns of losses (the diversification principle).

Are you saying with this that i can not use any combinations/patterns i want measuring Imbalance/Ecart/STD, because then i will at some times face very extreme events that we can not handle.
But what does (the diversification principal) do that norrow down does very extreme events not to happen?
I ask as i don't understand how you can reduce random fluctation and variance using (the diversification principal)!!!

The significance of RTM is not so much that it predicts winners but rather that it limits losses and keeps stakes low (which should be the prime directive).

Are you saying that if i wait for one sequense to hit 3+ (Ecart/std) then regression will clumb or hit more frequent when it shows towards the mean, so the path avoid you to face very long losing strings?
With other words that variance and fluctation will be lower then usally?

The logic is really very simple, although some will deny it because it smacks too much of the gambler's fallacy, although it isn't: Assuming the worst sequence from hell can run to 5 standard deviations (confirmed by me in millions of trials) and that such a sequence can come at any time - which would you rather have to cope with? 5 standard deviations of losses or at most 2 standard deviations? the latter, of course. Therefore, start your current play at 3+ standard deviations.

So you say there is a benchmark where you base your playing model upon 5+ ...?
(Same as Marigny claim, worst reckorded imblanace is 5.45 STD)
So with other words we will not during our life time see or break a world reckord and will always face a window or random string that will not go beyond 2.0 STD during our play from 3+ ... is that correct?

But then still i could face some serios loses as i don't know in which order a random flow of 2.0 STD comes with - all i know is that 12 events contra 2 events is 2.0 STD.
This means that even if i start from 3+ so could worse cenario be 12 loses in a row.

Then if some one would use the Gap Mehtodology by Holloway with even money bets and find that the largest gap is 22 during 1 milion trails.
If this person start betting from 10 he or she would face the same probability, or am i wrong here.
I know we talk about extreme events and the bell curve limits and that the random flow comes in waves, but what would be significant different?

The typical objection is that being independent trials, it doesn't matter what the deviation is before you start playing, none of what has gone before counts for anything. According to these math dogmatists it's just as likely that you will encounter 5 standard deviations on top of the current 3 sd giving 8 sd, and so on to an infinitely receding horizon of standard deviations like the pot of gold at the end of the rainbow which you never actually get any closer to. I prefer common sense and more importantly, experience.

So you say the random flow has limits and we can set a benchmark and play "cut point methodology" against very extreme events?

Add some notes to my reply above:
A window of 12 contra 2 is 2.5 STD - but if you start from 3.00 and add 12 contra 2 you end up with 4.01 STD
I think and belive the main idea is to divide the window into several windows of attacks as the random flow reach the benchmalk 5.45 STD
For example if you divide the window into several windows you can betting three times several times to prevent the final cut point.
Does sequenses or windows determine how smooth or agressive you have to bet.

[AsymBacGuy]

@Bayes.

Hi!
I'm a big fan of MdG so I perfectly know what you mean. His theories were applied at roulette with disastrous results eventually the author went in misery, mostly because he tried to look for deviations created artificially.
A detailed software has demonstrated that MdG main system (waiting for 3sf deviations and then bet after etc..) at roulette doesn't work at all. If interested I could send you a copy.
So mathematicians seem to be totally right: there's no point of intervention in trying to take advantage of the "ecartes" (deviations).

That's what happens at roulette, a perfect random machine.

At baccarat things significantly change for the reasons I stressed along my post.
So, IMO the RTM (regression effect) along with other parameters could play an important role in determining the possible profitable betting spots.

Anyway I strongly think that without utilizing some MdG ideas, it's literally impossible for a player to find out a FB edge and to make it working in practice.

as. 

   

[Bayes]

@ Patrik,

Yes, to all your questions.  You're an expert in cut-point methodology, do you no longer think it's effective? and if so, why not?

Regarding diversification, it's just a fancy way of saying "don't put all your eggs in one basket". A  simple example in roulette which FLAT posted would be to bet 3 units on an even chance. There are lots of ways of doing it, but instead of putting 3 units on red,  you would put one unit on each of 3 sixlines, the first being the "coldest", the second the "hottest", and the 3rd "average". The variance is reduced compared to betting only on red, even though the return is the same.

@ As,

Yes there are problems with using "artificially" generated deviations, but I don't use RTM as an exclusive method, it's a total package including MM, and I'm not aiming for flat betting either, just a way to avoid the sequences from hell. That's enough for me, and it really does work.

[Sputnik]

Regarding diversification, it's just a fancy way of saying "don't put all your eggs in one basket". A  simple example in roulette which FLAT posted would be to bet 3 units on an even chance. There are lots of ways of doing it, but instead of putting 3 units on red,  you would put one unit on each of 3 sixlines, the first being the "coldest", the second the "hottest", and the 3rd "average". The variance is reduced compared to betting only on red, even though the return is the same.

Bayes i think i understand diversification, assume i have 2.0 window after 3+, then i can divide the window into three sequential attacks and still cover the limit of 5.0.
But i like practical exampels:

Lets say Red has the value of 1 and Black has the value of 1.
Now i have 14 Reds and 2 Blacks.

R R R R R R R B R R R R R R R B

1) Now if i start betting now i will chasing for regression as its not present and the Ecart 3.00 can grow stronger (here i can face 17 loses before i reach Ecart of 5.04)
2) But if i wait for one more black to show before betting the Ecart drop to 2.5 (this means i aim for lower values and regression towards 2.32/2.13).
Assume we aim to hit only once.

So which way should one take?

Assume i divide attacks into three attempts:
Then the Ecart not get so much effect with up and downs (strenght) if i pick option 2 above.

For example using diversification: (as i understand it)

3.00 X B 

3.15 L R
3.29 L R
3.44 L R

3.13 X B

3.27 L R
3.41 L R
3.54 L R

3.26 X B

3.40 L R
3.53 L R
3.65 L R

We end up hovering around 3.00 and 3.50 using diversification (and this is about hitting once).
Maybe i think wrong, but i find it hard how diversifictation would work when you aim to catch regression (if we talk about winning one bet only).

I don't see a solution for option 1, so did not discuss it any further.
Also diversification might work if you aim to win several bets of regression towards the mean, but then it get more complex.

[AsymBacGuy]

@ As,

Yes there are problems with using "artificially" generated deviations, but I don't use RTM as an exclusive method,

I agree. Imo, RTM only isn't sufficient to build a long term winning plan.

Congratulations for your good results!


as. 

[Bayes]

@ sputnik,

Over the years you've posted many times on the CPM, for example:

http://betselection.cc/gambling-library/the-key-to-the-final-solution-marigny-de-grilleau/msg23664/#msg23664

Have you now come to the conclusion that it doesn't work? or perhaps just isn't practical because you have to wait a long time for opportunities?

Your example of:

R R R R R R R B R R R R R R R B

although it meets the 3 sd criteria, it isn't a long enough sequence. You should have 3+ sd over at least 20-25 decisions. Anything can happen over the short term, but over more trials is increasingly less likely. The tricky part is to get the balance right. Too few trials and you run the risk of continuing losses, too many and there is the chance that you will get bunched losses because you have to bet more spins. Also, In my method, I distinguish between differing distributions of losses, not just the number of losses in a sequence. For example, I look for minimum gaps between losses, and how long this can continue. I'll be going into detail on this on the web site. Then there's MM, including "damage limitation" which can be selected according to the particular pattern of losses you're targeting. In spirit the method is Marigny, but more fine-grained. I suspect that it's very similar to the late Martin Blakey's system.

P.S. The web site I was using seems to be unreliable, so I'm switching to google blogger. I'll post the new link in my signature soon.

[Sputnik]

@ sputnik,

Over the years you've posted many times on the CPM, for example:

http://betselection.cc/gambling-library/the-key-to-the-final-solution-marigny-de-grilleau/msg23664/#msg23664

Hi Bayes, yes that is correct that i know all about CPM/RTM, but i did not think about probability window with limits, that's new and we can add inspiration from Gap Methodology.
So i see that you come with frech ideas and improvments (diversification) that's great and that is the reason i would like to have this discussion.

There is many new things and new words that stand for something and i like practical exampels.
As above using diversification and you end up hovering around 0.5 STD and its also apply to Gap Methodology.
Where i cut the probability window into sequenses of three.
And my direct opinion is that it seems hard to apply such working principals.

You talk about a window of 2.0 or else we will experience a new world reckord, so the bell curve has limits during our life time playing, i like that very much, to base principals upon that.
So i am both for what you write and in the same time aginst what you write.
My problem is that i can not add your principals to a working strategy or at least i don't see any practical exampel to follow or learn from.

1) Do we play to win once or do we to win several times?

2) How would one use gaps with a probability window of 2.0 (17 events)
Do you get a equliberum towards same principal as card counting with positive expectation?
As we deal with determine amount of trails and some events are due to happen, because we have never reach beyond 5.04 STD.
Where the 3+ give us 2.0 and we know that we will get at least three events in our favour as it has to drop before reaching 5.04 STD.

Abstract:

If i have 3+ and a limit of 5.04 STD then i operate with a window of 2.0
Then during this path or probability window of example 17 events it has to be certain event to make the Ecart get weaker so we don't reach 5.04 STD.
Now i can caculate how many events that has to come to prevent 5.04 STD and does should be my benchmark for diversification and gap methodology.

I need to force my self into the window of probability to catch the same advantage a card counter has to avoid the distribution to hovering its way out of my betting strategy.

Testing gaps THEORY

If i pick 18/2 Ecart 3.57
Then the gap is 4 that increase each step by 0.5 Ecart
So 3 gaps of 4 and we reach 4.94 STD

3.57 our trigger (example)

3.70
3.83
3.96
4.08

4.20
4.31
4.42
4.53

4.64
4.74
4.84
4.94

Now the intressting part here is that for each step we cover a 4 gap and get a hit
Then we incrase the line with one more gap towards 4.94 STD

Conclusion:
This means that our expectation is at least two wins or more.
It has to happen or else the distribution will hovering around 3.57 and 4.94 with at least 1 or 2 hits - less is not possibal - because then we would go beyond 5.0 STD/Ecart ....

[Rolex-Watch]

Lastly, why oh why, one might ask, why the singles and threes concentration?

TBL, my friends, TBL. Time Before Last, or Decision Before Last, if you prefer.  It's nemesis is most easily definable:  the twos (2's).  Yeah...those "terrible twos", as they're often referred to.

Not so "terrible" for me, however, as I do very well to "side-step" them.  How?  By utilization of my "no-bet" option....in other words, I sit out those times when the 2's are besting the 1's and/or the 3's.

So I know my nemesis, I recognize my nemesis, and I react to my nemesis by "standing down".

So you side-step the 2's, and only playing for the chops and 3 streaks which you hope to capture  Absolutely amazing, darrn, never realised the game was so easy, golden in fact, utterly amazing..

Stay well, wishing you the best of it, my friend, blah blah