I've been looking at various posts and ratings and all types of free ratings and preposts etc etc.

Something that strikes me, is that over thousands of races, they tend to all end up at the same place. Of course there are differences in order and rated prices, but the end result appears to be very similar.

For example Timeform top two is the same strike rate as first and second favourites, as are free premier ratings.
Neither of them have any information about the dollar return, or the overlays.
This seems to be common with nearly all ratings.

Surely the test of any accurate ratings method is lay value in the overbet horses to rated price, and back value in the overlays to rated price.

I'm yet to see any ratings anywhere, that can beat the final betfair price for accuracy.
Reason being, that ratings cannot incorporate what we don't know.

Most of the time the winners and value if you take top fluc are from those under the rated price.

So are we back to systems again, which identify value regardless of market support?

Any ideas welcome...

__________________
RaceCensus - powerful system testing software.
Now with over 412,000 Metropolitan, Provincial and Country races!
http://www.propun.com.au/horse_raci...ng_systems.html
*RaceCensus now updated to 31/12/2024
Video overview of RaceCensus here:
http://www.youtube.com/watch?v=W821YP_b0Pg

Its well known that crowd theory produces the best prices. (ie the many always come up with the right price rather then the experts).

But I do agree with you the best test of a ratings method is to compare its overlay/underlay results for laying and betting.

If you can find an accurate rating that consistently outperforms with your laying/betting on the underlay/overlays then you have a license to print money.

Finding those ratings though are very hard to do. And then betting them through the losing runs is even harder.

I thought I would also add that the performance of a ratings method should be considered on its lower prices selections. Anything rated over $20 is hardly worth looking at because its only a 5% chance of winning. If you could get $40 which is a 100% overlay in price but is equivalent to a 2.5% chance. Your essentialy only getting an overlay of 2.5%. This is crazy to bet and is where laying/backing overlays comes unstuck.

At the same time if you take a $2.00 rated horse and you can get $3.00 then you have rated the selection at 50% chance and you have got odds equivalent to a 33% chance. This is a 17% overlay. Thats a lot better in terms of percentage overlay.

I would personally stick to gaining at least a 10% or higher overlay in percetnage terms. Bit for this you need to think chances of winning vs prices which is contrary to how many people think.


MInd you I am not e x p e r t in this and these are just my thoughts.

My experience with ratings is that the more exposed form there is, the more accurate the ratings.

Using ratings on a maiden with 12 runners and 7 first starters is ludicrous. Expecting ratings to work on every race is just not going to happen.

I have managed to whittle down the races I assess on average to 3 out of 8 on a race card and from there I only back a commodity if I believe the horse is value. That's why I believe Shaun is heading in the right direction with his ratings, he just needs to keep assessing his selections and bet when he feels he has an edge. Easier said than done.

I agree Chrome Prince that you're not going to beat the betfair final price or any bookie price for accuracy all the time, however if you tread carefully and selectively and edge can be found. There are still hype horses and horses that get spruiked week after week and disappoint as favourite.

Uselessbettor your points on overlays are extremely interesting and I shall investigate further. Also the point regarding losing runs is a fair one, mental strength and patience are just as important in this game. Not too many people can handle the rollercoaster ride that a 10% strike rate entails, no matter how profitable.

So to sum up, my thoughts on ratings:
- Choose races with lots of exposed form (most horses with 6+ starts and not many first uppers in the first 5 lines of betting)
- Bet only when you feel you have an edge (comes with experience)
- Once you've found your selection ask yourself, is this horse ready to hit its peak rating or progressive enough to go beyond its peak rating? (VERY important IMO)

The Schmile

We seem to have very similar approaches to this UB.

For example, one has to also think about the practicality or viability of backing the longer prices.
However, if you backed all horses greater than 500/1 before the jump on betfair, you'd make a profit every year.
It's just not practical though, unless it's a hobby or a bit of fun for you.

__________________
RaceCensus - powerful system testing software.
Now with over 412,000 Metropolitan, Provincial and Country races!
http://www.propun.com.au/horse_raci...ng_systems.html
*RaceCensus now updated to 31/12/2024
Video overview of RaceCensus here:
http://www.youtube.com/watch?v=W821YP_b0Pg

Thanks The Schmile for the contribution.

I've actually been looking at another angle of trials results, not to pick winners, but to try predict market movements.
That way, I have some hope of either selecting the best between Top Fluc or Betfair SP, or trading out.

As you say, being selective and judgemental is probably the key.

These are the results of one of my methods:


If I could predict which way the market would go, I'd be making a good profit.

__________________
RaceCensus - powerful system testing software.
Now with over 412,000 Metropolitan, Provincial and Country races!
http://www.propun.com.au/horse_raci...ng_systems.html
*RaceCensus now updated to 31/12/2024
Video overview of RaceCensus here:
http://www.youtube.com/watch?v=W821YP_b0Pg

Best of luck with the market prediction angle Chrome Prince!

I have been intrigued in recent weeks by your comments on winning without form and thank you for sharing your thoughts and experience on rating and pricing races.

The Schmile

Hey Chrome,

I believe its a combination of both, you have to start with the ratings and then systematicly elimintate poor races and under performing top raters. If your'e betting anywhere over half of the races on any given day you probably havent filtered enough.

From a systematic filtering of the ratings an overlay angle would be like skimming the cream from the top. its the place id suggest people head.

Cheers

Value Bet Method

Target 4 runners a race with your accessed prices,
say Neurals or Don Scott.

Bet 1 unit on the unders.

Bet 1.3 units on the overs.

Only target races where the Fav is 2.20+ in the live market.
Odds-on, look on.

Fields 8-13 runners. (Important)

It works a treat if our 4 horses get up on the day.

Backing the unders helps knock the edge of total loss.

The SR is approx 65%

If one bet all 4 horses at level stakes, same amount, it only shows a loss of approx -5% LOT
Which is less than just taking the fav to win.

So if we can tweak the value slightly our way , then we should be in profit.

__________________
Cheers.

I found this interesting and applied it to my selected races yesterday.

BR2 Fav under 2 no bet
BR6 fav under 2 no bet
BR7 14 runners no bet (winner was equal 5th rated 1.3x14.91)
*PR3 4.6 out 0 in
PR5 15 runners no bet (winner was equal 5th rated 1.3 x 12.45)
PR6 16 runners no bet (winner 5th rated 1 x 10.00)
*PR7 4.9 out 32.51 in 4th rated overs
SR6 16 runners no bet

So 2 races to bet on for an after coms return of 21.33

I made note of the other races.