INPACT VALUES

[formula_2002]

Impact values revisited.

IV =( win % of a particular horse) / the % of those particular horses in the entire assembly of horses

I’ve never been a fan on the IV ratings.
What advantage do I have if I know a horse has an IV of lets say 2.5.?

[Tom]

It tells me that that particualr factor is one not shared as many horse overall as by winners. If 85% of all winners raced in the last 14 days, that sounds boss. But is 990% of all horses ran in the last 14 days, you don't have anything special. But if 25% of all winners have the top last race Beyer, but only 11% of all horses have that figure, then you have something positive.

[formula_2002]

Tom, you only have something positive if those 11% of the horses returned a profit.

Thats just the problem when isolating on an IV..It tell syou nothing about profits or losses..

The more meaningful statement is the ratio of actual wins to expected wins ratio.
Any thing above a 1 is a profit, any thing below a 1 is a loss.

[Tom]

Few single factors will show a profit.
What I am looking for are factors that positively impact a race's outcome and those that either don't are are negative. I don't handicap with one factor, I use several in conjunction. The more "public" acceptable factors I can forget the better my chances of hitting better prices.
In HTR, K1-K2 factor hroses win 50% of all races overall. I cannot just dissmiss these horses from my handicapping - I don't always bet them, but I always give them a close look.
I am less strict about layoffs from 30-60 days now that I have done a layoff study. This used to be a reason to throw out a horse for me.

[JimG]

Originally posted by formula_2002
Impact values revisited.

IV =( win % of a particular horse) / the % of those particular horses in the entire assembly of horses

I’ve never been a fan on the IV ratings.
What advantage do I have if I know a horse has an IV of lets say 2.5.?

The IV factor wins 2.5 times the average for that slot.

[formula_2002]

Jim G.
So how do you make money with that information?

[JimG]

Originally posted by formula_2002
Jim G.
So how do you make money with that information?

If your really interested I would make 2 suggestions.

1) Go to www.frandsen.com and ask them for their "White Paper" on creating profiles for All-Ways. For all I know, you may be able to download it from the site.

2)Get a copy of Mike Nunnamaker's book Modern Impact Values. His web site used to be www.minnowbear.com

Jim

[formula_2002]

Jim, i've been using All-Ways for three years and have a data base of over 100,000 horses. I'm very familar with the "White paper" you suggested.

I'll check out the other site.

Thanks
Joe M

[formula_2002]

Jim I just did a quick read of the other site. Once again they speak of win% rather then profit..

I'm e-mailing a question, elaborated by a table .

I tried to attach it here but it's and excel file and canot be attcahed.

[Fastracehorse]

What is a K1-K2 factor horse??

fffastt

[Fastracehorse]

What do U use All-ways and "white paper" for.

fffastt

[formula_2002]

Jim


"Michael E. Nunamaker's selection service performance statistics"

indiactes a 9% loss in over 50,000 races

What am I missing here??

regards

Joe M

[Whirlaway]

Originally posted by formula_2002
Jim


"Michael E. Nunamaker's selection service performance statistics"

indiactes a 9% loss in over 50,000 races

What am I missing here??



Well, Nunamaker is an honest man, at least by horseracing standards. I'd think it would be impossible to achieve break even results while making your selections without knowing the odds and handicapping every race. A 9% loss is highly respectable.

Nunamaker's Modern Impact Values, on the other hand, is mostly worthless, and outdated to boot. Save your money.

[Rick]

Nunamaker's performance statistics are interesting because they show a reasonable limit on win % that one could expect from playing every race. Of course, there are some tracks where one could do a little better, mostly higher class tracks or those with short fields, but his results are for ALL tracks.

His 9% loss is due to using too many factors in his equations. When you attempt to maximize win % you always run into this problem. If you reduce the number of factors to a reasonable number, say three, the ROI is much better without sacrificing much in win %.

[formula_2002]

FROM www.minnowbear.com

"This section details how Michael Nunamaker's selections have performed over a sample of over 50,000 races."

This does not imply that the sample includes "every" race.
Races may very well have been selected races.

[formula_2002]

IF YOU THINK A 9% loss is good, with out much effort, I just ran a program on my data base of 12,822 races and which produced a 10% loss over 6705 races at a win % of 33%.

But you will not make bank deposits with that kind of "good"

[BillW]

formula_2002,

I'm trying to understand your point and am failing miserably.

An impact value is a piece of data, a fact. I don't understand how one can like or dislike a fact.

Is it your interpretation of that fact that you dislike?

Bill

[formula_2002]

Bill, i'm implying that an impact value is has no ...value .
Yes it is a fact. But "winning horses have four legs" is also a fact.

How do you make money with either of those two facts?

[Rick]

I think you'll find that Nunamaker explains in another place that his win % was higher before he started doing it for every track. It should be easy to beat his win % and ROI by just using selected tracks and using as little as two factors. My results on all races are about 29% wins and a little better than break-even. But, what's really important is that the factors I use find true overlays in almost half of those races.

Impact value is just another way of expressing win % and not too useful on it's own. But, if you can find factors that have a decent win % (or impact value) and an ROI better than average, you may be able to combine them and come up with something that has a high win % and good ROI. The trick is to find relatively independent factors. Most combinations will actually result in a worse ROI because they are correlated with one another. In a situation like that, the second choice will usually show a better ROI than the first choice.

[Tom]

Originally posted by formula_2002
Bill, i'm implying that an impact value is has no ...value .
Yes it is a fact. But "winning horses have four legs" is also a fact.

How do you make money with either of those two facts?

Don't bet three-legged horses <G>

To answer your K-1,K-2 question, K is a rating factor in HTR software. K1 is the number one ranked horse, and K2 is the number two ranked horse. K3 is third, ....
K1 wins an average of 30% and K2 wins about 20% generally speaking - it varies depending on the type of race. Knowing that K1-K2 win 50% of the races, they are involved in half of the top hole of exactas. You have to respect these horses even if you don't bet them. K1-K4 win around 75% of all races.

ITEM Plays Wins WN% $ROI WP%
----------------------------------------------
K1 04186 1215 29% 0.86 48%
K2 04188 0798 19% 0.80 37%
K3 04184 0633 15% 0.81 31%
K4 04188 0511 12% 0.85 25%
K5 04172 0335 08% 0.72 19%

Races Tested: 4186

These results are for all tracks, all races, the last month, about.

[Tom]

Originally posted by formula_2002

"Michael E. Nunamaker's selection service performance statistics"
indiactes a 9% loss in over 50,000 races

What am I missing here??

regards
Joe M [/B]

Profits!

[hurrikane]

I think you may be missing the point a little on impact value.

Impact value is used to asses the value of any given factor on the data to see if it performs better than would be expected and therefore 'perhaps' a bias or advantage exists.

IV is not used to find positive ROI. In fact the ROI is usually useless in IV because many times it includes factors (ie runstyle) that will have more than one horse in a race.

You may have 50% of the horses in all races are either 'E' or 'F' style horses. Lets say 80% of the races are won by 'E' or 'F' horses.
You have an early speed bias at the track. Will betting every E and F horse in a race produce a profit..well..lets not get crazy.

But..if you check a track and see that 50% of the races are won by early speed horses you may think that you have an early speed bias. Maybe not...not if 80% of the horses running in the races are early speed. You may actually have a bias exactly opposite of what you believe is happening.

It happens everyday ...even on this board. Someone sees three races won by the #1 horse and thinks..'there's an inside bias'.
They are assigning an Impace Value based on the fact that the #1 horse should not win 3 races in a row. Therefore there must be a bias. (which I believe is wrong 90% of the time)

It has nothing to do with ROI or profits.

[CapperLou]

Hurricane:

You've got it right--you are on the right track!!!! I'll mention what are some impact values to me as follows:

In no particular order:

1. MSW dropping into MdClaimer for 1st time.

2. Horse dropping 25% in claiming price from last race.

3. A single 4 yr old maiden in with all 3 yr old maidens early in season; that is in January to April in particular.

These are just a few well known Impact Value situations. You have to look for these and get an overlay if possible--folks--they will happen--just look and wait for them.

Now, if you want to use All-Ways software and their files--you will find that they will show IV situations for a particular track after you have a database of at least 100 races & preferably more of course. You really need 200 to see something.

If you do not want to spend the kind of money necessary for the above--then follow some of the 3 suggestions I oulined above and play them only when they are overlays. You should do just fine based on my past experience. Hope this helps!!!

Happy Holidays!!!

CapperLou

[Fastracehorse]

I cannot disprove U're IV's but I hate most droppes.

Especially into Mdn claimers.

I actually luv risers with Mdns.

fffastt

[formula_2002]

Hurrican and Capperlou.

We know what Impact values IS NOT. In what proven way can it help you on the road to profits.
(Lets skip it's function in explaing that 100% of the four legged horses win 100% of the races!!)

As far as All-ways, 200 races will not substanciate any thing.

You really need several thousand races in EACH odds range to come away with reliable factor predictions.


Every factor you can think of, including the three Capperlou mentions, must be valued by reviewing thousands of races in EACH odds range.

B. Fabricans book "Horse Sense" does an excellent job explaning just how many races must be reviewed in oder to establish various confidence levels.

If only it were easy...

[CapperLou]

Formula 2002:

You are so right----it's NOT EASY!!!!

I also agree with you about thousands of races where All-Ways is concerned, generally speaking. What I was talking about previously where All-Ways was concerned was looking for recent patterns or Impact Values; which it does fairly well. Trends and patterns are constantly changing and it's up to us to spot it manually or use a software product that will spot it for us.

The three spot impact value methods I mentioned before are time tested and they can be filtered to get an even better ROI.

This game is very complex of course and it takes extremely hard work and a good mind and most of all after that is instincts--and that takes many years to obtain--there are a few people who master this quickly--within 5 to 10 years, but for those others who finally become profitable it can take decades!!!

Again; Who Said this Was Easy????????

CapperLou

[Tom]

Whille you are out gathering data from thousands of racces, I am going to betting on what is happening right now. This year, Del shippers are a good bet at AQI. Bet them now, this year. If you look at data from the last 20 years, you might find that DEL shippers are throw outs at AQI. Play the current events untils they stop winning. Now is how.

[cj]

Tom,

One of the best pieces of advice I've seen here! Bravo!

CJ

[CapperLou]

Tom:

That's why I mentioned current trends above. I did very well with DEL shippers to Saratoga---and it continues---you have to spot trends---current ones to me are more important than past ones in most cases.

CapperLou

[cj]

Another example...

The rail at Hol used to be death for a horse sprinting, but it has come back from the dead the past year. Using long term stats would have missed many a nice winner.

CJ