Neural Settings
 

After the discussion within the last couple pages of the lay the outsiders thread I'm pretty keen to try adjusting the Neural Settings (i.e. 0,1,2,3,4,5) based on impact values. This may be in relation to race prizemoney, distance, venue, day, state and conditions (Basically the detals in relation to the race and not the horses themselves).

However my issue is that my record keeping is similar to that of a caveman as I only have good knowledge of excel, which unfortunately (to my knowledge) isn't so useful as a database tool when your trying to record a number of variables for say a 12 horse field.

So two things
a) Does anyone have a database of the R&S Neural's recorded (preferably on setting 5) along with the other race details as mentioned above
b) Does anyone profficient in setting up a database to record such things wish to give me a hand in getting started?

I don't mind doing all the data mining as I actually enjoy messing around with numbers, hence why I continue to build and adapt on my own ratings method. Any outcomes of interest that I find i'm obviously happy to share with whoever can provide some help or possibly the forums as a whole depending on how useful the outcomes are.

If I can't find anyone interested i'll have to try go it alone, cheers.

I have them on default settings. (3 for everything but wet. Wet = 0)
I also have most of the other stuff depending on the time frame you want it for.

But if someone else has the data can they please provide it as I am currently inundated with a heavy workload and I don't have a lot of time to provide this in a useable format at the moment.

UB, default settings would still do the trick, only makes a difference on the decimal places a bit. No rush on the data either, not a project I plan to have done by sunday or something, just an idea I think has some merit and deserves some investigation.

Let me know if/when you got time and we can possibly work on this together and see what we come up with.

Please contact me , I will be able to help you in some way
feebles36@gmail.com

Managed to get my own spreadsheet together and have been recording the neurals now since about friday last week. Obviously going to need a LOT more data before the Impact Value's start being representative of the true impact of their variables but at least its a start.

At the moment i'm gauging the impact the race variables have on the neurals to hopefully get a more accurate adjusted version of the neurals. I gave it a try yesterday with the limited data and had some success plus the Adelaide Cup winner became the 4th top rater instead of about 7th or 8th in the neurals.

Race variables im calculating IV's for:
- Distance
- Prizemoney
- Conditions (G,D,S,H)
- Venue (Tatts code's at the moment, with more data will likely move to Metro/Provincal/Country the same as garyf)
- Race Rating
- Race Age restrictions
- Race Sex resitrictions
- No. of Runners

I might start posting some results in another 2-3 weeks time once the data has been given some time to at least build up a little. I still need to get in contact with Feeble though and see if I can do it on some historical data as well. Will post more updates as I go for those interested.

Hey all,

Another update regarding my trial of adjusting Neural settings based on impact values of the race conditions as mentioned in previous post. I've now got 500 races recorded over the last 10 or so days. A lot of the impact values are starting to take shape and on saturday I used the adjusted Neurals to back some horses before my basketball grand final in the afternoon. Managed to hit 4 winners with odds of $2, $8, $9 and $12. Not sure if it was just a lucky morning of punting or the impact values are really working their magic.

There are still some racing elements that simply haven't had enough races recorded yet so their impact values are way over/under stated (For example Mares only races). These type of races will likely take a couple of months at least before their impact values are representative of the actual performance.

However once I have 1,000 races recorded sometime in the next 2 weeks i'll start collecting the data on the adjusted Neurals as well and their performance to see if it does improve the performance of the Neural ratings. Initial thoughts is that it does improve the results and that it can possibly used both from a backing point of view and laying point of view. An idea i've got at the moment is to lay the bottom rated horse in races of 8 or less horses, bottom 3 in 12 or less runners and bottom 5 in 13+ horse fields however i'll have to run some tests obviously to see if its viable.

Anyway just thought i'd keep people up to date and will start getting some results up in the near future.

thanks for the update. Looks like an interesting approach.

Almost finished collecting the 1,000 races of data to start looking at the viability of my adjusted Neurals. On a side note I used them again yesterday while I was home sick from work and jagged a $17 winner with my 2nd bet, I'm very keen to see how it performs long term.

In the meantime i've been reading a few of the threads on Neurals lately (mainly michaelg's) and i'm wondering if anybody has a database of the Neurals and are willing to test something for me a bit longer term.

I've tested it myself over the 950 races i've got recorded so far but would like it tested a bit more extensively. The rules are extremely simple and based on my testing the average odds required for it to break even is about $18 (Lay system). If anyone could help me out it would be greatly appreciated, obviously if it proves to be seriously profitable then whoever helps me out can use it.

Will post more updates regarding my adjusted Neurals in the coming weeks.

Noorals
 

Onya Eva. I'm watching with interest. I saw where somebody posted about the boring "noorals" but I find them fascinating as an area on which many approaches could be buiilt - both backing and laying.

If it uses the basic settings and doesn't use the price field then I can help out longer term. Otherwise I'm no help.

Sent you an email UB,

Will definitely keep you posted Speedy, I find the Neurals interesting as well. I've tried the whole do your own ratings thing but since I don't have a database I think its just pointless as i'm just basically guessing what factors to weight higher and by the time I do trial and error and see what works it could take me a decade. At least with the Neurals they are already done for you and all you have to do is simply adjust them to suit your needs or to what you believe yields the best result.

I wish I had a database because i'm a real numbers guy and love trial and error with facts and figures. Considering I work in financial planning thats no real surprise but I almost enjoy the thrill of the chase of finding a holy grail (something I can consistently bet to make me some bonus money on the side) rather then actually achieving it.

Anyhow it will be about another 2-3 weeks before I can give any kind of indication as to how my process is going but will certainly keep you all posted.

neural's
 

hi all,
i am a newby on this site,i love neurals.
i have a system.
this are my selections for racing on saturday 30th march
oakbank race 2----number 2=apres
oakbank race 7----number 2=avoid lightning
rosehill--race 6----number 2=it's a dundeel

good luck.

can you tell me if bhagwan is still about ,no posts lately
cheers .
bartle.

evajb001,

I tried replying several times to the email but internode keeps rejecting it. Here is my response to your system:

=============


Are you laying these or backing them ?

I tested it over 4 random months from 2012. Of those months only 1 month
showed a small profit of $2. The other months all lost considerably using
tatts figures but I didn't test betfair prices. Over the 4 months tested there was a return of $780.90 from 1111 selections. I didn't go test other months as it was losing and I assumed you were betting the selections.

==============

Hi UB,

Not sure why thats the case, maybe my inbox is too full? Anyway just thought i'd let you know its a lay system not a back system, can you re-test for me please?

Thanks

I have tested 4 months again (same 4 months) using betfair prices for laying.

Unfortunately it showed a loss just over 10% after commission. The increased prices on betfair always make it hard. For instance there was 1 winner at $810 on betfair which was under $100 on tatts.

I tested this using both methods of laying (straight $1 or to lose $100). Both laying staking methods showed a similar loss.

As a guide if your using tote odds for testing then you need to be under 50% returned if your method uses odds any higher then $20. I also tested your system under $20 but it showed an even larger loss but the odds range over $20 also showed a loss by itself.

I couldn't find a way to make it profitable.

Thanks UB for that reply, ill track myself as well shortly just for interest sake.

I decided to back my top adjusted neurals rater today using half kelly staking with the winning probability set to an average of my assigned winning % based on neurals and the winning % of the neurals top rater from all previous recorded events in races with the same or similar details. The following are the stats:

30 bets today 9 (30%) winners 17 (56%) placings
Average Winning Odds = $4.70
Average Place Odds = $1.84

Winners - Flat Stakes
Outlay $300
Profit $123
POT 41.00%

Winners - Half Kelly
Outlay $280.14
Profit $90.10
POT 32.16%

Placing - Flat Stakes
Outlay $300
Profit $30.50
POT 10.17%

Placing - Half Kelly
Outlay $280.14
Profit $1.20
POT 2.41%

Anyway just thought i'd post the results, guess flat stakes worked the best, wonder what the result would've been if I did some kind of chasing.

http://www.************************....au/form-guide/

Try the neurals here,,right hand side of page

oh dear it's not allowed..

There are other neurals beside r/s

I'm sure another neural site would be allowed to be shown here. Its not like pro-punter is in competition with neural sites. Just write out the name of the acual site without the www and dot com dot au parts and we can find it for ourselves.

punters paradise,,click on form guides,,select a race,,scroll down right side of page..

Okay I think they MAY BE in direct competition. :oops:

evajb001,

I sent you Reply All email with a snippet of VBA code to fix to that workbook's slowness.
If my suggestions don't make sense on where to put the code, let me know.

LG this includes you to



Thanks.

Thanks Puntz

Cheers LG

I forgot to add "grinning" at the end of last post

Thanks Puntz, will take a look tonight.

I don't use the neurals anymore but the change to code may help speed up the other spreadsheet I use.

Cheers

evajb001,



yes, it'll universally benefit.

Wow new site to me syllabus23, thanks mate!
Once loved doing the neurals, will have to check them out on PP next week.On wet tracks though they have less consistency but could be very useful on drier going in a couple of months.