Percentages & Probabilities 2012 - "Review"

[jdhanover]

I thought it was worth separating this out to a separate thread from Dave's pre-release thread so that folks can more easily find and comment, ask questions etc about Dave new publication, Percentages & Probabilities 2012.

I got the P&P-2012 last evening...and as someone else posted, couldn't put it down. It is a tremendous piece of work that I think will benefit virtually all players. It is very rich in data content and Dave provides insights and overviews that help guide the reader through the information.

As one should expect it is not a cookie cutter 'here is the path to winning at the races system' but rather it is chock full of information and insights that new and advanced handicappers can learn from. He explains away certain myths and how certain 'truths' have changed over the years (a couple have been turned upside-down!). And he gives the reader a view on how to approach the game that is likely to be different from what you are currently doing.

Dave does include a number of 'systems' that, as he says, can be the basis of a robust handicapping approach. [Some may work stand-alone btw, I don't know - it would take a lot of testing to figure out. My guess is these are foundational and the user will need to adjust to his/her desires and goals]. He includes data and worksheets for HSH users, Excel users and for paper-and-pencil users.

Understand the book is very data intensive (I will need to go thru t several times to pull out all of the information that I want to get from it) but even if you aren't completely a numbers person, there is a lot to help your game.

My take, in summary, is that this is a fantastic addition to any handicapper's library that will at a minimum provoke new thinking that the user can incorporate into their approach.

Dave - great work...and thanks!!!

P.S. In the spirit of 'full disclosure' - I am a HSH user (about 9 months); prior to that I was a paper-and-pencil handicapper

[goforgin]

Basic Rating question. If both Horse A and Horse B have a speed rating of 85, do they each get a "1", and then the 3rd horse with a speed rating of 82 gets a "3"?
Also, let's say the top 3 speed ratings are 85, 84 and 82. Very close. Is it always ranked 1, 2 and 3 regardless of the closeness of he numbers?

[jdhanover]

85/85/82 is 1, 1, 3

I think regardless of the closeness you order them, so 85/84/82 is 1, 2, 3 (but will look for Dave's confirmation)

[JimG]

Since I do not have a large database, I thought it was worth the money spent. It is very intensive and will take several reads through to really absorb.

It is a large scale effort and was worth the cost to me.

Jim

[Dave Schwartz]

Goforgin,

Jon nailed it.

Basic Rating question. If both Horse A and Horse B have a speed rating of 85, do they each get a "1", and then the 3rd horse with a speed rating of 82 gets a "3"?

That is correct. I used Fred Davis' description of ranks:

1st - None are better
2nd - 1 is better
3rd - 2 are better

This kind of covers it pretty cleanly.

85 - none are better = 1st
85 - none are better = 1st
82 - 2 are better = 3rd


Also, let's say the top 3 speed ratings are 85, 84 and 82. Very close. Is it always ranked 1, 2 and 3 regardless of the closeness of he numbers?

Correct. Such is the problem with ordinal factors: it ignores the differences (sometimes referred to as "gaps").

Unfortunately, building a workable and simple system demands ordinal factors rather than raw. (It is the "simple" part that is difficult.

When you get a chance, take a look at the "Above Strength" factors as well as "Difference from the Top." You will find a lot of meat there.



Dave

[Tom]

I like the Davis format of things...I used that methodology for years so it is familiar. Lots of reading to do....LOTS of reading!

Looks like I picked the wrong week to give up speed!

[valueguy]

This is very good info.But unless you can computorize the methods in some way
you will burn out very quickly trying to use the systems as a stand alone
platform.I figure it would take 3 hours to do a 10 race card.
There are all knds of ways of using Dave,s info to augment your own handicapping methods .I love the ways to look at speed for different races.
This book will be the new standard for impact and probibility factor handicapping.

Good stuff Dave!!!

[Dave Schwartz]

I think you will find it is more like 12 minutes per race to start, then you get it down to about 5-6 minutes.

Such is the nature of handicapping. If anybody thinks they are going to not use a computer and handicap a race in like 2 minutes, they are kidding themselves.

Well, they can certainly make picks in 2 minutes but you won't make strong picks.

[Tom]

Dave, this is a MASSIVE amount of data!
Kudos for your hard work.

One question I have is form - can't a horse be a 1 and 2 at the same time, or do you just take 1 as the priority over 2?

As far as the systems go, what I did with the Davis data back in the day was to set up a spreadsheet for each one in used and just filled it all in then had it calculate the ratings and odds.

[valueguy]

I get your drift Tom.Set up a spread sheet so as your factor determinates
become data input.All i would have to do is expand the P&P worksheet to get
the results i need.

Thanks much.

[LottaKash]

I
Well, they can certainly make picks in 2 minutes but you won't make strong picks.

Maybe ?....:cool:

best,

[goforgin]

Dave, again, very good work. Question about pulling the trigger on a wager? How do you decide which races to play? Curious if you handicap 3 or more tracks and then only play select races? Play only when the hit rate and the $NET are high? I created a few test spreadsheets and played around with the odds, wgt-score, etc. I guess in short, if the spreadsheet outputs a wager, trust the numbers and it's always a play, or wait for races with higher $NET's, focus on FH's?

[Dave Schwartz]

I handicap all the races then look for value.

(When I say, "Look for value" I mean in the race, not in a single horse.)

One of my struggles these days is that I am only finding value in about 70% of the races. In about half of those races I make a significant wager (determined by Kelly), spread across 1,2,3, or even 4 horses. In the other half of those races I make a less significant wager.

The "large" wagers, those below the "average" bet size are marginally profitable after rebate, while the above average bets are solidly profitable.

Truly, it has taken me many years to appreciate the power of Kelly.

[Dave Schwartz]

Oh, and as for your question about "how many tracks:" for me the game is lots of fun up to about 4-5 tracks. After that, it becomes too much work.

[rowboat]

HAVE YOU EVER RESEARCHED COMBING SYSTEMS AT SAME TIME FOR SYNERGY

[Dave Schwartz]

From that idea was the concept of "Monty Hall" born.

[classhandicapper]

Correct. Such is the problem with ordinal factors: it ignores the differences (sometimes referred to as "gaps").

Unfortunately, building a workable and simple system demands ordinal factors rather than raw. (It is the "simple" part that is difficult.

Dave

I saw this approach used in the late 70s in a book by a couple of professors that created a regression model for racing. I still have it somewhere in my library. They only looked at ranks, not gaps. That always struck me as an issue, but I didn't have the stats skills or energy to try to improve on it. That's why I always wound up operating on feel instead of a model (not that "feel" doesn't have issues also).

[Tom]

So, how is everyone doing with this stuff?
I have used a couple of the "systems" so far - CS, CR, Cnw2, Stakes......with positive results, but still small samples.

[jdhanover]

Looking at ways to incorporate the info into what I am using. Not sure it works as pure 'stand-alone' (Dave never said it was), but the ideas in it have sparked some new thoughts for me. Its rather powerful stuff. I just need to figure out how to best use it.

[headhawg]

At the risk of sounding like a fanboy, I really enjoy Dave's products a lot. They are well-researched and always packed full of information. So while anxiously awaiting the Renegade Handicapper {waiting...waiting...waiting... :D }, I decided to pick up P & P. I knew it would be quality, but it turned out to be better than I expected. I mean, there's a lot there. A lot.

My review echoes what jdhanover and others have posted. P&P fits me well because, as I understand the importance of db research/datamining, it's not something that I prefer to do. (I also don't have a giant database to work with so that's a little problematic as well.) Dave did the bulk of the work in P&P so now I can focus on putting the findings to good use. I am going to try to automate some of the systems so that it will easier to track profitability. Hopefully that won't take me too long. (Way too much working and not enough "playing". <sigh>) I'll try to post back what I find, and maybe Tom and others can update us now that they have larger samples.

[jk3521]

Any updates on the profitability of any of the "systems" included in the P&P
package?

[Capper Al]

I like the book and all the data. I was intimidated by it's size at first. Found after I had it printed, reading through it was easy. Read it three times over after that.