Money management

[Secretariat]

I was wondering how some here are handling money management.

Theoretically, I like Mitchell's idea of establishing a line, AND betting overlays. A few problems with this though. 1) Requires a very accurate line 2) What's the cut off on maximum win percent horses to cover (i.e. an overlay who has a 2% chance to win doesn't make me comfortable with the runouts) 3) This is a big one. One HAS to have access to the near closing odds to employ this approach. I have difficulty being able to constantly watch, etc. Is there any online betting where you can make a bet depending on the odds? For example, could you place a bet "if" the horse say hit 4/1 at 30 seconds to post? You can do this on the market, but don't know of any horse betting sites that do this.

Second, Dave Scwartz's HMI is a great approach, as is his opponent method, but these schemes don't adjust for the complexity of a race. For example, say you determine Race 3 is my pick of the day, but I like Race 7, but not as much. These schemes treat each race as if they are the same in the betting. I know beyer did a prime bet and playing bet. Curious how others have done this. For example, some races may lend themselves to exotics, and some to a pure win bet. In allotting a wager, does anyone here ADJUST their wager based on the strength of the "assurance" factor of the race at hand. If so, how do you factor it into your overall bank?

Just curious on others ideas on this.

[kingfin66]

I have been playing grass races almost exclusively. My approach is very simple:

* Rate the horses and identify the contenders. For grass races, I will limit my contenders to 4, unless there are ties. I have a method for doing this, but it isn't on topic with your question.

* I enter the "live" odds for my contenders into my spreadsheet, and if the live odds are bigger than my odds for the horse, it is a bet. So my thingy might look like this:

Horse 1 92.9 2/1 3.7/1 No
Horse 2 92.5 2/1 4.1/1 No
Horse 3 89.9 5/1 7.9/1 No

Incidentally, these ratings, etc. come from today's 7th race at Aqueduct, and illustrates the sometimes sad part of Value Betting. Horse 3, was actually the #4 horse that won the race. It was my 3rd and last contender, but the two short-priced horses on top both had very strong odds lines. I would have had to have a better price to have a bet. But, that's racing...

* A couple of things worth noting. I am looking at the 1st two overlays from among my four (or less) contenders. I also want price, not just value. So I am wanting 4/1 odds or better, but usually get much better.

* Dave's HMI is a great tool. I guess I don't really follow your point regarding prime plays vs. action plays. They are all the same to me...a bet is a bet is a bet. I believe that this is Dave's theory as well and that he has said that he pretty much plays every race and bets multiple horses in every race that he plays.

* I prefer to play unit wagers for my win bets and use HMI for my exactas. Most people would likely not like how I am playing exactas. I am simply boxing my contenders, up to five. I have to be flexible with HMI using this type of betting because HMI might tell me bet $28 on a race. If my exacta box is $12 for horses, I need to either bet $24, or go up to $36.

* Wouldn't that be great to be able to put Buy orders in on horse races. I seem to recall this being discussed somewhere before.

[twindouble]

I was wondering how some here are handling money management.

Theoretically, I like Mitchell's idea of establishing a line, AND betting overlays. A few problems with this though. 1) Requires a very accurate line 2) What's the cut off on maximum win percent horses to cover (i.e. an overlay who has a 2% chance to win doesn't make me comfortable with the runouts) 3) This is a big one. One HAS to have access to the near closing odds to employ this approach. I have difficulty being able to constantly watch, etc. Is there any online betting where you can make a bet depending on the odds? For example, could you place a bet "if" the horse say hit 4/1 at 30 seconds to post? You can do this on the market, but don't know of any horse betting sites that do this.

Second, Dave Scwartz's HMI is a great approach, as is his opponent method, but these schemes don't adjust for the complexity of a race. For example, say you determine Race 3 is my pick of the day, but I like Race 7, but not as much. These schemes treat each race as if they are the same in the betting. I know beyer did a prime bet and playing bet. Curious how others have done this. For example, some races may lend themselves to exotics, and some to a pure win bet. In allotting a wager, does anyone here ADJUST their wager based on the strength of the "assurance" factor of the race at hand. If so, how do you factor it into your overall bank?

Just curious on others ideas on this.

Sometimes I wonder if I just waste my time and breath here on this forum. When it comes to this post, it's about to cement my thinking. :bang:


Good luck!

[Dave Schwartz]

Sec,

Personally, I do not view "bet structuring" within a given race as money management. I view it as, well,... bet sturcturing. <G>

Perhaps some of the guys that have attened the Basics of Winning Classes would care to comment on what we apply there. That is the entire purpose of that class.

(If nobody else comes forward, I will discuss it a little, but I would prefer it to come from someone who is applying what they have learned.)


Dave

[Secretariat]

Sec,

Personally, I do not view "bet structuring" within a given race as money management. I view it as, well,... bet sturcturing. <G>

Perhaps some of the guys that have attened the Basics of Winning Classes would care to comment on what we apply there. That is the entire purpose of that class.

(If nobody else comes forward, I will discuss it a little, but I would prefer it to come from someone who is applying what they have learned.)


Dave

Thanks Dave.

I guess it goes to king's comment as well:

He states:

"I have to be flexible with HMI using this type of betting because HMI might tell me bet $28 on a race. If my exacta box is $12 for horses, I need to either bet $24, or go up to $36."

Using his example, say HMI say the bet should be 28, and my bet structure goes to 24. I've 4 dollars to structure into my bet.

Additionally, what if I "rate" my race bet strength on these type of races at say a normal ROI of say 1.40 and, another race at say 1.28. HMI does not adjust for varying success rates among races, but seems to look at an overall success percentage. Any suggestions on how to adjust for this? Should one adjust the HMI percentage per race strength? In other words, do you advise altering the percentage on a race success level, other than an overall level to take advantage of those situations where you excel?



Thanks.

[Valuist]

To answer your question what should the cut off be for the odds you anticipate: I say if you think the horse has less than a 10% chance of winning, forget about them in the win slot. If one of these horses is 25-1 or so, I say look to key them underneath in either exactas and/or tris.

[Dave Schwartz]

Sec,

When someone uses variable bet sizes, HMI should be used to determine what the normal bet should be. Then you can go up or down from that as you decide.

Example...

Suppose your basic approach is a 2-horse, 60/40 psuedo-dutch and a $2 exacta box with your top 3. In other words, your "standard" bet is $22.

#1= $6w
#2= $4w
$2 box (1-2-3)
$22 total


Think of that as a 2-unit bet.



The next full step is a 3-unit bet:

#1= $9w
#2= $6w
$3 box (1-2-3)
$33 total


If HMI calls for a $28 wager you are still betting only $22. If (when?) HMI calls for a wager of $33 or more you step up.

There are many other creative ways to apply HMI's increasing or decreasing. For example, you could increase the win bets before the exactas. For that matter you could even increase the wager on #1.

The point is that there is always a way to address this. You just have to think of a workable strategy.


Another thing to do is that you don't have to adjust the bankroll after each race. So, suppose you adjust every third race? If HMI calls for a $28 bet (and your next even unit would be $32) think of it as borrowing $4 from the next two wagers. If you get to the third wager and don't have enough for a full wager, just adjust the BR early, putting back the money you did not have time to wager.

At our HMI seminar in November we are actually going to work through some of these more creative examples.

I'll tell you where some of the most creative ones come from... sports bettors! They not only write straight tickets but also parlays and it gets very complicated, especially if they parlay from Saturday to Sunday but want to adjust on Sunday morning (before some of the wagers have been completed)!

The point is that you can always make it work. You just need to define a strategy.


Dave

[finfan]

Dave,

I'm basically a spot play player. All of my spot plays generally fall into three categories: low hit rate (<20%), medium (20-30%) and high (>30%) with varying ROI's. Should I have a separate HMI bankroll with different parameters for each level, or should I lump them all together? Just wondering your thoughts on this. Thanks

[Secretariat]

Thanks Dave and Valuist for good ideas.

One final question on your example Dave. Your answers have got the cobwebs here spinning.

Using your example:

Example...

"Suppose your basic approach is a 2-horse, 60/40 psuedo-dutch and a $2 exacta box with your top 3. In other words, your "standard" bet is $22.

#1= $6w
#2= $4w
$2 box (1-2-3)
$22 total

Think of that as a 2-unit bet.

The next full step is a 3-unit bet:

#1= $9w
#2= $6w
$3 box (1-2-3)
$33 total

If HMI calls for a $28 wager you are still betting only $22. If (when?) HMI calls for a wager of $33 or more you step up."

...

Let's assume on my first wager it is an Allowance n2l race, and my personal records indicate that they represent these are my best race types to play. The second wager is a Maiden Claiming race and my records indicate they are also profitable, but not nearly as good as the Allowance n2l.

In the above example, I am using a fixed overall bet percentage computed on my overall success rate, and unfortunately in the above example, am actually wagering MORE on my LESS successful type of race.

I know nothing is fixed in stone, but what I'm wondering is could one "borrow" from a future wager, but it would have to be "paid back" on a lesser one to balance the ledger within HMI? In other words could I add a column called prime loans where "special optimal situations -best bets" could be tapped into provided they are paid back from the profits. If one lost, that loan would have to be paid off out of future "lower bets" or "profits" on future hits.

Anyway, Dave you've got me thinking. Thanks much.

[PriceAnProbability]

I was wondering how some here are handling money management.

Theoretically, I like Mitchell's idea of establishing a line, AND betting overlays. A few problems with this though. 1) Requires a very accurate line 2) What's the cut off on maximum win percent horses to cover (i.e. an overlay who has a 2% chance to win doesn't make me comfortable with the runouts) 3) This is a big one. One HAS to have access to the near closing odds to employ this approach. I have difficulty being able to constantly watch, etc. Is there any online betting where you can make a bet depending on the odds? For example, could you place a bet "if" the horse say hit 4/1 at 30 seconds to post? You can do this on the market, but don't know of any horse betting sites that do this.

Second, Dave Scwartz's HMI is a great approach, as is his opponent method, but these schemes don't adjust for the complexity of a race. For example, say you determine Race 3 is my pick of the day, but I like Race 7, but not as much. These schemes treat each race as if they are the same in the betting. I know beyer did a prime bet and playing bet. Curious how others have done this. For example, some races may lend themselves to exotics, and some to a pure win bet. In allotting a wager, does anyone here ADJUST their wager based on the strength of the "assurance" factor of the race at hand. If so, how do you factor it into your overall bank?

Just curious on others ideas on this.

I use the "fire at will" method, and if that fails, the "Ahab" method of chasing my losses. Like in Star Trek II, when Kahn's shields were down, and Kirk said "Fire!....Fire!!....Fire!!" and destroyed the ship.

This works because the power ratings I make have shown a profitable ROI in just about every category, so as long as I bet the value horses, I know I will come out on top in the long run. If I lose too much money too quickly I get very conservative and play in the "penalty box" until I win my way back out.

Generally, if you are truly good enough to show a flat-bet profit in the win pool, highly aggressive wagering (in win and exotics) will pay off.

[Dave Schwartz]

I'm basically a spot play player. All of my spot plays generally fall into three categories: low hit rate (<20%), medium (20-30%) and high (>30%) with varying ROI's. Should I have a separate HMI bankroll with different parameters for each level, or should I lump them all together? Just wondering your thoughts on this. Thanks

No need for separate bankrolls, just separate strategies.

There are just so many ways to structure this. Perhaps you could have 3 strategies... each one addressing the potential Kelly of the strategy. For example, the high hit rate strategy might call for a wager that is 3 or 4 times the wager size of the low-hit strategy.

Of course, you've got to work out what is best but somewhere there is a strategy that will work for you.


Dave

[Dave Schwartz]

Sec,

In your example, (like FinFan's) you would probably be better off with multiple strategies. Try this...

Simply grade each bet for hit rate and probability, assigning a numeric value to each one:

High = 4
Med = 2
Low = 1

Then multiply them together:

Hit profit units
High High 4 4 16
High Med 4 2 8
High Low 4 1 4
Med High 2 4 8
Med Med 2 2 4
Med Low 2 1 2
Low High 1 4 4
Low Med 1 2 2
Low Low 1 1 1


Make it as complex or as simple as you wish. The point is that the highest profit/hit rate wagers should have more money bet on them that the low profit/hit wagers.


Hope this helps.


Dave

[andicap]

Dave,
I have two different type of plays, the WP bets that hit at a decent rate and my exotics --exactas mostly -- which are much more streaky because of stricter odds demands. (Lower odds horses are more overbet in exactas/tris and I'd rather try and hook them up in pik 3 where the payoffs can be generous with a 3-1/5-1/4-1 horse result.)

Question: To use session play effectively what should your minimum hit rate be on my WP plays? I use a 3-1 minimum for the most part (5-2 in smaller fields or for really exceptional plays).
In other words, if I'm hitting 20% win at an average of roughly $11, say, vs. 35% WP at an average of a bit over $6 (plus rebates) -- roughly 10% ROI before rebates --- is either one better for using sessions?

[Dave Schwartz]

Question: To use session play effectively what should your minimum hit rate be on my WP plays? I use a 3-1 minimum for the most part (5-2 in smaller fields or for really exceptional plays).

In other words, if I'm hitting 20% win at an average of roughly $11, say, vs. 35% WP at an average of a bit over $6 (plus rebates) -- roughly 10% ROI before rebates --- is either one better for using sessions?


In my opinion, almost everyone should use session play. About the only exceptions are those players whose bets are always maxed relative to the bankrolls. An example of such would be a rebate whale. (And even they could use sessions but it has no impact on how they wager.)

Sessions are just a logical way to break your play down into "games" so that you can develop a "session win pct" and easily "take profit" on a periodic basis.

Sessions also allow you to play more agressively without fear of going broke.

However, I think you have read this article:
http://www.horsestreet.com/freestuff/articles/OffCenter/index.html

... and are thinking of a session as having a flat number of plays as Howard Sartin used to suggest. Personally, I prefer the "goal session" approach where you play until a session is won (i.e. a goal is reached) or lost.

Example: Suppose you have a $1,500 bankroll. You could call a "session" 30% of the bankroll and play until it triples or goes away. In other words, your session would be $450 and you would play until the $450 reaches $1,350 or goes away.

When it reaches $1,350 you then combine the session back into the bankroll (making the BR now $2,400). The next session would be 30% of $2,400 or $720.

The beauty of this kind of approach is found in the fact that you eventually develop a "session hit rate." )Hit rates of 60-70% or more are quite common for the profitable player.) Once you have a "session hit rate" you can easily compute the optimum session size ("Kelly").

If your hit rate was only 60% at 2/1 then you would find the optimum session percentage was 40% of bankroll.

As you know, I am not fond of using the Kelly Criterion on individual wager sizing simply because you don't really know what your hit rate or advantage is. But session play (as described here) has a rigid hit rate and a very precise percentage.

Speaking of creativity (from my previous post), I spoke to a guy once who told me he uses a fix number of wagers (i.e. "session") as a single bet in HMI. (huh?)

That is, he uses HMI to size his betting units for his next 10 races. At the end of the 10 races he writes up is profit or loss in HMI and sets the bet size for the next 10 races. So, his "standard unit size" (which has nothing to do with Randy Johnson) might be $7 in the first session and, after adjusting for a big winner change to $8 or $9 in the next session.

Just remember that the more consistency you add to the mix the better HMI will work.

For those players who experience wild fluctuations caused by big hits and long droughts, stretching the HMI adjustment over more bets helps even things out.


Dave

[kenwoodallpromos]

I'm just surprised there is not a concensus on the issue. How many methods work?

[classhandicapper]

Theoretcially, you should bet more on high probability/high ROE horses
and less on low probability/low ROE horses.

My real world experience does not coincide with that theory though.

Regardless of whether I make an odd lines or rank the horses in order of preference and only play when the odds appear to be an overlay, my results are always the same. My high probability horses win more often, but my long term ROE is always higher on the my lower probability live horses at longer odds.

I think a fundamental reality of the game is that we are all working with partial information and understanding. Sometimes we are even working with inaccurate data and a misunderstanding of what it means. That makes it extremely difficult to determine how much of an edge we actually have on any given race even if we are pretty sure we have an edge.

This all argues for flattish bets of a size that you can cope with both emotionally and financially. Flat bets also eliminate a lot of time and energy that might be better spent structuring the wager rather than deciding how much to bet on the race.

I have categories of bets but remain flat within the category.

Prime, Action, Place Wagers

[Dave Schwartz]

I'm just surprised there is not a concensus on the issue. How many methods work?

All of them? None of them? The issue is what is right for you.

The issue really is that you must be able to show a flat bet profit. If you can do that then money management can improve your consistency and ROI. It just cannot turn a losing approach into a winner.


Dave

[Secretariat]

Dave,

Exceptional advice. Thanks so much for really insightful ideas.

Sec

[PriceAnProbability]

I'm just surprised there is not a concensus on the issue. How many methods work?

That depends on where you are as a handicapper.

First off, *none* of these methods will work if you can't pick the horses to get the edge, so don't forget that part of the equation. People are so eager to quit their day jobs that they'll take any evidence of profitability as proof that they are now "professionals" and attempt to design "professional" money management. If they hit a bad streak, they then have to decide if it's just a bad run, or if maybe they weren't as profitable as they thought. A lot of guys who make money now are no more certain to make it for the next year or five years than anyone else, while others have tried and true methods that aren't leaking to the public.

Generally, if you are still worried about money management, your handicapping could use an upgrade. Can you hit a cold exacta or trifecta, or superfecta? I do that regularly. How about a cold dd, p3 or p4? I do that regularly as well, sometimes for ridiculously high payoffs. How do you do with big tickets? If you key one horse over four in the tri, does your singleton usually win and does another horse slip in the money? Only when your tickets can consistently do what you want them to do, or come close, should you worry about how to manage your money. Once you have an edge, the only question is do you play it safe or not.

When you reach the point that you can evaluate every horse in every race accurately and quickly, and have full confidence not to second-guess yourself, and you learn basic ticket structuring, you show a profit in all categories for one year or longer, you'll be prepared for just about any situation that strikes you, and your main concern will be how much to get down, or maybe if you kill your price or not.

Bet distribution is not money management, and most of your questions seem directed at that end of the equation. For money management, if you want to be a true warrior, and you have proven to yourself that you are profitable in the long term, then you should be willing to fight down to your last dollar or up to your first million, whichever comes first. If you fail, find another line of work and keep this as a hobby.

[sjk]

your main concern will be how much to get down, or maybe if you kill your price or not.



If you play Kelly with an advantage and have the temperament to increase your bet sizes as prescribed (which I do not) pool size becomes a major factor.

How do you (or others) integrate these considerations into your action?

[PriceAnProbability]

If you play Kelly with an advantage and have the temperament to increase your bet sizes as prescribed (which I do not) pool size becomes a major factor.

How do you (or others) integrate these considerations into your action?

I was serious when I said I use "fire at will" money management. I see something I like, I bet it. Pretty simple. I have various temporary stop-limits that I use, but I generally don't think much about it. My focus is on finding good horses.

If I lose, I use the "Ahab" money management strategy and chase my losses until I recover them, getting more conservative with each chase, or what I call the "penalty box."

[sjk]

I was serious when I said I use "fire at will" money management. I see something I like, I bet it. Pretty simple. I have various temporary stop-limits that I use, but I generally don't think much about it. My focus is on finding good horses.

If I lose, I use the "Ahab" money management strategy and chase my losses until I recover them, getting more conservative with each chase, or what I call the "penalty box."

I was serious too. I don't use any stop limits or worry about chasing losses because losing never enters into my thinking.

Do you have a response as to how to integrate Kelly with pool size restictions?

[PriceAnProbability]

I was serious too. I don't use any stop limits or worry about chasing losses because losing never enters into my thinking.

Do you have a response as to how to integrate Kelly with pool size restictions?

Just check the pool sizes in the charts.

[sjk]

Knowing the pool size is important but that was not what I was interested to learn about.

I hear about people earning $1 million a year and I wonder what their strategy is for dealing with the pool sizes.

At least for the smaller tracks you cannot scale success with smaller bets without significantly impacting your ROI. For example if you are used to cashing a $200 exacta 5 times in a $50,000 pool and these types of plays are important to your overall ROI, playing that ticket 50 times is going to take away much or all of your advantage.

Do bigger players just ignore the smaller tracks because they are not worth the trouble?

The obvious answer is to get a good rebate because rebates are scalable when ROI is not.

Wondering what strategy would be appropriate to the player who relies on ROI and enjoys playing many tracks.

[Dave Schwartz]

One million per year? Try fifty million per year.

Actually, I believe that Mitchell addressed that in one of his books but I could be wrong. It is on my personal agenda for the near future.

Dave

[sjk]

Dave,

I would be interested to know what you come up with. I have no plans to play at that level but am curious if there is a standard strategy for working around the pool size problem.

SK

[Dave Schwartz]

I believe a model is not too difficult to create, but it is not without effort.

First, you capture the tote at the time when you generally make your wagers. (Both individual wagers and pool totals.) Then you capture it after the wagering has stopped (usually about the time the trophy is being presented <G>).

Next, you determine the percentage of money coming down after you place your wagers. This becomes a "pool multiplier" of sorts.

This is about where my plan gets fuzzy but it would seem the next step is to determine how different your wagers typically are from the late money. (This task would be more difficult.)

Finally, your model should include the impact of your wagers on the pool.

It is, of course, a best guess because you have no idea if the horses you are wagering on are the same ones the whales are wagering on.

[speculus]

Thanks to Dave and PrinceAnProbability, this thread is really turning out to be a good one on Money Management, although the original poster was perhaps more interested in knowing about BETTING STRATEGY than MONEY MANAGEMENT.

What would be the correct definitions that would help in not confusing the two?

Betting Strategy: The structuring of bets for a race AFTER handicapping a race and knowing the market odds

Money Management: Simply, deciding how much money to risk on the next bet

Are these definitions ok? Or they need to be refined further?

[sjk]

Finally, your model should include the impact of your wagers on the pool.



It is this part that I was curious about an answer to. Suppose you are making a bet into a pool that you expect to be a certain size and you have an expectation of a known strike rate and odds. If you bet a very small amount the odds will not be affected; if you bet to much you will clearly kill your ROI advantage.

There must be a simple formula that shows the maximal expected dollar return and the bet size that should be made to achieve this. The larger bettors must use this sort of formula all the time. If there is a rebate involved this would play into the solution.

Have never taken the time to work it out since I don't bet large enough amounts to kill my own prices but would be curious to know the answer.

Lets say you set a goal of making $1m and plan to do this by betting 5,000 races. You need to have a positive expectation of $200/race to achieve the goal. Is it possible to make a bet with a positive $200 expectation in a $20,000 pool? My guess is that unless you have a very strong ROI (or rebates) it may not be possible.

[betovernetcapper]

I've toyed with the Kelly Criterion for years while it always looks great on paper, there are a lot of problems in the real world.
Kelly uses three factors to create bet size-prob' of winning-tote board odds-size of bank roll. For an example, if my BrisBet account has $1000 and my morning lines and tote odds on A-B-C are:
# ML Odds TOTE
A .....2-1........4-1
B .....3-1........10-1
C .....4-1........3-1
The proper Kelly bets are
$239 on A
$207 on B
$82 on C
Putting pool size aside, this stratagy involves putting 50% of my bankroll at risk in one race. If I win a couple of races consecutivly, the stratagy would suggest betting 4-5 figure amounts on the next race. This is just too far outside my comfort zone.
What I've found useful is to use the Kelly Criterion against my largest normal bet. For me a large but doable bet is $50 and that's what I use as my base.
So example bets become
A=$12
B=$10
C=$4
Not exciting but that's what I do.

[Lefty]

I've toyed with the Kelly Criterion for years while it always looks great on paper, there are a lot of problems in the real world.
Kelly uses three factors to create bet size-prob' of winning-tote board odds-size of bank roll. For an example, if my BrisBet account has $1000 and my morning lines and tote odds on A-B-C are:
# ML Odds TOTE
A .....2-1........4-1
B .....3-1........10-1
C .....4-1........3-1
The proper Kelly bets are
$239 on A
$207 on B
$82 on C
Putting pool size aside, this stratagy involves putting 50% of my bankroll at risk in one race. If I win a couple of races consecutivly, the stratagy would suggest betting 4-5 figure amounts on the next race. This is just too far outside my comfort zone.
What I've found useful is to use the Kelly Criterion against my largest normal bet. For me a large but doable bet is $50 and that's what I use as my base.
So example bets become
A=$12
B=$10
C=$4
Not exciting but that's what I do.
You need to get Huey Mahl's papers on the Kelly. He had lots of variations for racing and sports. He says never bet over 8% of br any situation regardless of what Kelly says it should be.
For racing he says you should refigure everything every 15 races are so as win pctg and avg odds do not stay constant.

[Overlay]

I've toyed with the Kelly Criterion for years while it always looks great on paper, there are a lot of problems in the real world.
Kelly uses three factors to create bet size-prob' of winning-tote board odds-size of bank roll. For an example, if my BrisBet account has $1000 and my morning lines and tote odds on A-B-C are:
# ML Odds TOTE
A .....2-1........4-1
B .....3-1........10-1
C .....4-1........3-1
The proper Kelly bets are
$239 on A
$207 on B
$82 on C


If full-size Kelly wagers are beyond someone's bankroll or comfort level, one other approach might be a "fractional Kelly", where wagers are made in the same proportion, but risking only a constant fraction or percentage (such as one-half or one-quarter) of the full Kelly amount.

[Dave Schwartz]

The issue with applying Kelly the way it has been described in these last few posts is that the examples all use the projected odds as the probability as opposed to the real probability.

Two problems:

First is that just because you make a horse 2/1 does not mean he really has a 33% chance of winning.

Second, even if you do your homework as much as possible and are able to say conclusively, "My 2/1 shots win at a rate of 29%," that does not mean this 2/1 horse has a 29% chance of winning. It means that collectively they win 29%.

Of course, the degree to which your probability is accurate will be directly reflected in your profitability over time.


We believe that wagering is a two-phased process: handicapping (where one makes probabilities) and wagering (where one exploits the difference between our handicapping and the public's).


The biggest problem we face as probability makers is that we must find a way to figure the tote board into the handicapping phase as well as the exploitation phase. In our methodology we call that smoothing.

And guess what? It still does not make a truly accurate set of probabilities (How would you know if it did?). It does, however, make them significantly more accurate than they were.

BTW, for those interested, my Basics of Winning video project is falling behind a bit. (Okay - I haven't worked on it in 3 weeks and probably won't touch it again until after Thanksgiving.) So much so that I may do one all-day workshop before I get it finished. Stay tuned.


Regards,
Dave Schwartz

[PriceAnProbability]

First is that just because you make a horse 2/1 does not mean he really has a 33% chance of winning.

Most people don't grasp this, and just assume because some handicapping author told them to make a morning line, that they are actually any good at doing so. It took me seven years to refine my value line method to where I can rely on its percentages to hold up over time.


The biggest problem we face as probability makers is that we must find a way to figure the tote board into the handicapping phase as well as the exploitation phase. In our methodology we call that smoothing.[quote]

Only to the extent that tote information alters a horse's probability should it be included, as in it's additional accurate information to what the handicapper possesses.


[quote]And guess what? It still does not make a truly accurate set of probabilities (How would you know if it did?). BTW, for those interested, my Basics of Winning video project is falling behind a bit. (Okay - I haven't worked on it in 3 weeks and probably won't touch it again until after Thanksgiving.) So much so that I may do one all-day workshop before I get it finished. Stay tuned.

If you use a mechanical method for producing ratings and lines, you can model that method to where you can pretty much know what to expect results-wise.

Another thing I find is that one shouldn't even consider "spot playing" until they can play every race and win, because they won't educate themselves properly if they develop tunnel vision. Some of my best rating methods come in race types I would have dismissed as crapshoots, but by forcing myself to solve these puzzles, I gained (painful) experience that led to several breakthroughs.