Neural Data Help - Please?
 

Hi All

I'm working on something where the top rater has all the neural settings set = 0, except TA and JT that are set to default (3).

Last couple days have found winners @ $14, $11, $7, and $4 with a reasonable(?) strike rate.

I was wondering if anyone might have the last 3, 6 or 12mths worth of data with these settings?

I would be very appreciative and would be happy to share my findings - as I have in the past with Chrome, Wesmip1, Michaelg and others....

Contact me: stix_hotcopper at aanet dot com dot au

Some qualifiers today are:
Bendigo R1-1, R6-14...back with some more soon...




Thanks In Advance

Anyone? :eek:

I dabbled with that setting some time ago & what I found is that it went really well for some days, then fell into a total hole other days .

Very hot & cold.

I found the std settings worked as well as anything else.

With some very good paying winners.

Looking at the Neural forum , a lot of the chaps seem to compare their various settings with one another , on how well their top 2 selections performed as a combined percentage, so as to judge whether their settings had any legs or not.

Rather than just relying on their top selection only.

I feel if one wants to get the best out of the Neurals , I found its a good idea to target the top 2 of whatever favourite Neural setting one is using.

Then separate those top 2, using ones favoured separation method & go from there.

The separation method does not necessarily have to be all that logical, if it's shown to perform in the past, that way , the better payers can be snared.

Try & keep the separation process as simple as possible.
That way it has a greater chance of repeating itself into the future.

Thanks Bhags, I've got a method already which picks up selections using Neurals and 100 pt raters, which is doing well since I started in October last year - and thanks to some of Wesmips data he provided - but he's gone AWOL...

Often on a lazy Sunday will look at top two neural raters and "normalise" the CF and CP figures to get a selection. For example Top rater has 185 points (CF 80 CP 15) and 2nd Rater has 130 pts (CF 20 CP 25) - i look at the combined CF and CP points (95 vs 45) where there is a big discrepancy between two values - in this case CF points - I look at the 2nd rate and if it's CF was 80, would it out rate the original top rater, which it would in this case (130-20+80=210) then it becomes the selection.

This works pretty good and don't have too many races before you hit a winner, small staking plan until a winner hit and then either stop or start again...anyway again something to think about and play around with.

I was looking at TA and JA setting in conjuction with Unitab and got some decent priced winners, although Sunday was not the best day...

Anyway if someone has the TA & JA settings for any period of time, it would be appreciated!!

Be Good...

Thanks Stix, that looks interesting.

So one would take the combined CF+CP then add that to its original rating , is that right?
e.g.
1st rated say 180 + CF25 + CP45 = 250
2nd rated say 170 + CF 40 + CP 45= 255

The 2nd rated becomes our selection?

Sorry , I just re-read what you said.

I think I got the jist of it..

Example:

CP for Oratrice is 65 v Black Tycoon 10

If Black Tycoon's CP rated 65, its final rating would be 151.5 - replace the CP of 10 with 65 (96.5-10+65-151.5) - Black Tycoon's adjusted rate is 151.5 and is > Oratrice's 107.5....hence it becomes your selection....it works sometimes ;)

TAB	HORSE	NR	HCP	CP	CF	TIM	SCR	JA	TA	JT	BP	WET	CRS	D	$	DLR
9	Oratrice	107.5		65	3	0	0	0	0	8	0	0	0	10	6	16
5	Black Tycoon	96.5		10	21	0	0	0	0	18	11	0	0	5	16	16

Thanks for that .

I got it now.