STOP, before you comment.

[speculus]

Originally Posted by jfdinneen

Speculus,


As you know, when summarizing data distributions both a measure of central tendency (e.g., mean) and a measure of dispersion (e.g., standard deviation) are the minimum requirements.

In assessing handicapping performance, there is too much emphasis on expectation (i.e. edge) and too little on volatility (i.e. luck). However, very few handicappers make enough wagers in their whole lifetimes (e.g., > 100,000) for edge to outweigh volatility (or risk of ruin - loss of bankroll) as meaningful metrics of overall performance. As an extreme example, a punter on Betfair recently matched GBP 40 @ 1000/1 in-play on eventual winner, Tioga Gold. at Southwell, 13:20, Jan 02, 2007 - collecting GBP 40,000 in a single race!

Therefore, I would ask that when evaluating money management plans, all handicappers consider risk of ruin, volatility, and expectation (in that order) as critical metrics!

Best wishes,

John

Thank you John for pointing out some interesting and important points.

I am monitoring all the following factors for three systems: Flat Bet, HMI and Invest (temporarily I have christened the formula Invest. Not because it is an investment system but because it is under Investigation.

For your information, for all three, I am concentrating on the following points during this investigation under different weightage rules:

1. State of bankroll
2. ROI
3. Average of "growth of bankroll" (this is computed as totaling, in percentage terms, the growth of bankroll and dividing by the total number of bets in the sample)
4. Gambler's Ruin points
5. If no gambler's ruin takes place, then maximum percentage of capital used (this is calculated as the lowest "state of bankroll figure" for the sample plus the amount of the next bet as proposed by the system
6. Markup details for growth of capital in multiples of the starting capital
7. Biggest and smallest bets as % of "starting" and "current" bankroll
8. Volatility needs to be checked only in the case of HMI as it is the only one of the three for which sequence influences the state of bankroll. Right now I am not exactly concentrating on that as Dave Shwartz, the man who devised it, seems to know it like the back of his palm, and he can be called in at a later stage for his opinion provided the formula really stands up to all these tests. Why bother Dave if the investigation leads to the scrapping of the formula for not being what it promises at first glance or for not even being usable for most situations?

If you have something else to add, please let me know or write mail to me.

Thanks again for wishing best luck,

[speculus]

Originally Posted by spilparc

I always thought Bob Stupak, the former owner of Vegas World, was the first one to come up with the idea of 100 x odds on craps. [The following article says I don’t have it quite right. http://www.casinoplayer.com/archive/...100x_odds.html]

But for a while there you could get 100 x odds at Vegas World and the Horseshoe. (And other places as well.)

Receiving 100 x odds on the pass line lowers the house expectation to 0.021 percent.

“Mathematically, with 100-times odds, the casino's edge can be reduced to a piddling 0.02 percent; that's an expected loss of only 2 cents per one hundred dollars of action on the pass line and odds.”

A friend of mine who knew Stupak fairly well, quoted him as saying, “just give me a small fraction of a percent return, and I’ll eventually grind them out of all their money."

Steve

Those of you who are bringing games with FIXED NEGATIVE EXPECTATION (on account of constant probability and constant odds) please realise you may be doing harm to this thread. I am NOT claiming the formula can improve the ROI in those kind of games, but only in games like horse racing where EDGE and ODDS are VARIABLE.

"Can there exist a money management formula/method that can convert a negative edge to a positive roi? Something like a flat bet loss of 16% to a positive ROI of 20%?"

Yes in theory.
If you place more on the winning bets than on the losing ones.
Say your multiples of unit stake are selected at random each bet (the method). There will now be a few sequences of total profit purely by chance - the big bets have gone on winners and small bets on losers. There will also be far more sequences of increased loss (depending on the degree of fortune reversal required eg -16% to +20%) . That effect (in reverse - ie you can still lose) happens even if your edge is positive and you make "random" variations to the multiples on your unit stake.

[Light]

Spec

I don't understand your point of posting .If you found something that no one else has been able to do,then why would you post about it? If you want to discuss it then like D.S. says we cant really discuss something whose details are unknown. My personal intuition (excuse my frankness),is that you are a bit like my brother in my roulette story. Idealistic but a bit naive.

[PriceAnProbability]

Originally Posted by speculus

Can there exist a money management formula/method that can convert a negative edge to a positive roi? Something like a flat bet loss of 16% to a positive ROI of 20%?

Before you jump at my throat for asking the stupidest question, I urge you to please STOP before posting your comment, and go through my earlier posts at this forum to check how seriously I take my pursuit of handicapping and money management, and I hope you will at least agree I am not one to raise such questions without serious reason.

I value my time, and yours too.

What happened was this. After comparing my results against the backdrop of a money management formula that I was using, and Dave Shwartz's Horse Market Investing (HMI), I decided to switch to the logic of the HMI for this year's betting action (2007).

However, one thing I could not get comfortable about was the system's SUPERSENSITIVITY to the WIN-LOSS sequence. If the win-loss matrix changed, the numbers would change. I had always beleived a sound money management method, in principle, must imply that the end result (the figure of profit after the last bet in the sample is made) MUST STAY THE SAME whichever sequence the wins and losses took place. As you know, FLAT BETTING, the most reliable (although not the most lucrative) money management method, is practiced by many winners precisely because of this reason.

HMI, otherwise an excellent tool in many respects, does not satisfy this condition. That's when I dug out my old research (some 400 pages of unorganised hard work) and started looking at it. After spending more than a day with the material, what is staring me in the face right now is an ILLOGICAL (illogical, as per the general accepted wisdom on the subject at present) claim that says you can NEVER convert a negative edge into a positive ROI no matter what you do. It is impossible both theoretically and practically.

Now my question is: Do we believe it becuse we FEEL it is SELF EVIDENT? or has anyone ACTUALLY given a mathematical proof of the alleged truth? I learnt my early money management from Dick Mitchell, so I am still not able to believe the possibility although, as I said, it is staring me in the face (at present, in the form of too small a sample to be convinced, but who knows it might be possible to have huge samples reflecting the same ILLOGICAL truth?

I see a member of the MARTINGDALE family has shown up.

I don't study money management, but I used to study a lot of MATH, which is all any of this is. I would say at least take a college statistics course if you want a good grounding in money management (and making speed figures, which I did right after stats 101).

The pros use the "fire at will" stratgey because their handicapping is solid enough to justify it. If you can't make this work, become a stronger handicapper and don't quit your day job until you can make it work.

[Dave Schwartz]

Spec,

Volatility needs to be checked only in the case of HMI as it is the only one of the three for which sequence influences the state of bankroll.

You keep saying this but I still have not seen it.

HMI works from peak-to-peak. If you don't have peaks (and of, of course, valleys) then it doesn't work well. But then, without peaks and valleys why would you need ANY money management? - You are either rich or dead broke in a hurry.

I would contend that, if sequence is showing as so important in your HMI runs that you are trying to get HMI to do too much of the work; I would suggest there is a strong possibility that you are over-betting your edge.

Of course, this is pure speculation (no pun intended) on my part, having neither seen your samples or really being desirous of working hard to prove you are incorrect.

How large a sample are you running through HMI? Are the samples break-even or better? How many wagers does it typically take to complete (win or lose) a session? How many random runs have you done?


Regards,
Dave Schwartz

[Dave Schwartz]

PS: Someone who is well-read on the topic mentioned The Gambler's Fallacy. As I recall, that is a really good example of why winning with a negative expectancy is not supposed to work.

It is also discussed in Alan Wilson's book.

[44PACE]

Changing your betting amounts may alter your outcome a little but not enough to overcome a negative situation. Wether you won or lost your last race or number of races doesn't provide for a predictable win or loss in the next race. If it did it, has more to do with the mental status of the person betting i.e. if you are the type of person who falls apart after a loss then you may sabotage yourself in the next race.


The best way to overcome a negative game is to improve your game ( pick more winners, make fewer bad bets). Concentrate on your strenghts and KNOW your weaknesses. WE all have weaknesses eliminate races were these weaknesses happen to be.

[singunner]

I'm convinced he has ignored me in as much as he hasn't responded to any of my posts (even the logical ones). Perhaps it's my tone, but then, I often take the absolutely impossible trivially, precisely as Holmes prescribed.

[Secretariat]

Originally Posted by ryesteve

Depending on the sequence, isn't it also possible to make the sessions even more negative than they otherwise would have been?

Yes, it is. However, since one is using a plateau (or kind of a modified bluntschli) one is increasing stakes in the second or third session to recoup losses from the earlier session.

For example, as to Holloway's description of plateaus, say you lost 8 units in your first session, and upped your base stake and goal in the second session. If you recouped in the second session you return to the first level. If you went onto a third session you may make up the money here. If now you lose your third session, you can revert back to the earlier session.

Within sessions on the Oscar system you only rise up on a win, not a loss, and therefore long streaks do not kill you as they do in other methods. However chasing a permutation that despite all your maneuvers can occur especially with a negative flat overall return.

In other words using a rise up on a loss overall session, and a rise up on win only within session works OK for me in casino play. But on handicapping I work more on grading my bets, and modeling my success rate at my grade bets. I have been playing with HMI by Dave and using a lot of those ideas as well.

But nothing in betting is guaranteed. Ever.

Speculus, a session I once had over ten bets. L-L-L-L-L-L-L-L-L-L
Is there any way your method can turn this into a profit?

[Cesario!]

Originally Posted by singunner

I'm convinced he has ignored me in as much as he hasn't responded to any of my posts (even the logical ones). Perhaps it's my tone, but then, I often take the absolutely impossible trivially, precisely as Holmes prescribed.

I like this quote and wouldn't mind adding it to my repertoire. Is it Oliver Wendell or Sherlock?

Seth

[PriceAnProbability]

Originally Posted by Dave Schwartz

You keep saying this but I still have not seen it.

HMI works from peak-to-peak. If you don't have peaks (and of, of course, valleys) then it doesn't work well.

"Peak to peak?"

[BlueShoe]

Short answer to this topic;yes it can be done.When discussing minus expectation games the math boys overlook a very important point;these games do not behave the same as random generated numbers.You can run a perfectly programmed computer simulation,and yes,the method or methods tested will fail and lose money.However,in real life at the tables your method might be a steady winner.There are actually a few guys grinding out a profit at the tables in Nevada or AC on a fairly consistent basis.So how do they do it,if not using a "system" or "method"?Mentioned earlier was Barstows "Beat the Casino"Some of his ideas have real merit,although a bit of tweaking might be in order.Oscars Grind?Ainslie (famed racing author) wrote a 1979 book called "How to Gamble in a Casino" that presents a practical modification of this method using stop loses and stop gains.Horrible losing steaks on one side such as 20 straight reds or blacks or 20 passes or misses?Why not get on the streak instead of bucking it?Few of you will recognize the name,but the late Louis G. Holloway was a great believer of going with what is happening now,not with what is overdue.Old LGH was a master in money management methods,and presented many in his writings and how to use them wisely.Slump breaking,stop loses,progressions without retreat,fractions of due columns,and other topics he covered.Walk up to a crap table and pick pass or dont pass and then blindly play the system that your cousins barber claims to never lose with or start a progression on your favorite post position,and yes,you will surely lose.Money management methods attempt to do one thing;put more money on your winners than your losers.Most fail,but a very few,used wisely,can actually work over a long period of time.

[Cesario!]

Originally Posted by BlueShoe

Short answer to this topic;yes it can be done.When discussing minus expectation games the math boys overlook a very important point;these games do not behave the same as random generated numbers.You can run a perfectly programmed computer simulation,and yes,the method or methods tested will fail and lose money.However,in real life at the tables your method might be a steady winner.There are actually a few guys grinding out a profit at the tables in Nevada or AC on a fairly consistent basis.So how do they do it,if not using a "system" or "method"?Mentioned earlier was Barstows "Beat the Casino"Some of his ideas have real merit,although a bit of tweaking might be in order.Oscars Grind?Ainslie (famed racing author) wrote a 1979 book called "How to Gamble in a Casino" that presents a practical modification of this method using stop loses and stop gains.Horrible losing steaks on one side such as 20 straight reds or blacks or 20 passes or misses?Why not get on the streak instead of bucking it?Few of you will recognize the name,but the late Louis G. Holloway was a great believer of going with what is happening now,not with what is overdue.Old LGH was a master in money management methods,and presented many in his writings and how to use them wisely.Slump breaking,stop loses,progressions without retreat,fractions of due columns,and other topics he covered.Walk up to a crap table and pick pass or dont pass and then blindly play the system that your cousins barber claims to never lose with or start a progression on your favorite post position,and yes,you will surely lose.Money management methods attempt to do one thing;put more money on your winners than your losers.Most fail,but a very few,used wisely,can actually work over a long period of time.

Examples seem loaded with the gambler's fallacy. In fact, they're soaked in it. Does this law no longer apply?

[singunner]

I guess the bar for entry into the forums is set pretty low. "Get on the streak". If I weren't at work, I probably would have had a good hard laugh at that.

As for the quote, Sherlock Holmes was famous for his "When you eliminate the impossible, whatever remains, however improbable, must be the truth." He says it many different ways, but the "whatever remains, however improbable, must be the truth" is pretty constant. My claim was simply that people here are expressing their uniquely human quality of babbling on about the impossible in such a fashion as to assume that it is not so. As soon as you accept a single fallacy into your argument, it weakens the entire argument considerably.

When I work with lawyers, they almost always include a phrase along the lines of: "if any part of this agreement is held to be unlawful, it shall not affect any parts of the agreement which remains lawful". The thing is, if you don't stipulate that, a single illegal term will invalidate the entire agreement.