wrong settings

[garyf]

hi all just a quick message to all neural users on the r+s website that the distance and p/money ratings are wrong and have been that way for quite a while eg race 3 today at grafton nos =5 soul brother check its overall p/money then its rating the distance column also has numerous discrepencies i rang them thursday to notify them they were completely unaware of the mistakes till i pointed several out to them they said they would work on the issue but unfortunately were unable to give me a date when it would be fixed as i'm not sure how long this has been wrong for it may be wise for neural users to re-check their figures
cheers
garyf

[wesmip1]

Can you explain what you mean by wrong ?

Nuerals are complex rating systems and just because a horse has the top prizemoney does not mean it will be the top neural figure.

[garyf]

hi wesmip1
i agree neurals are a complex set of mathematical formulas the example i gave to them was ballarat last thursday my call was transferred to robert who i believe is their head (i.t man) when i showed the examples he agreed their was a problem in the read that was calculating those two particular settings and said he would look into it re my email to the forum i've been using the neurals now for 5 years and find them an invaluable source in identifying the main chances in a race just re-read my example at gosford how does a horse have 4 starts EARN NO P/MONEYyet be equal 2nd rated it's just a mistake that's all one that their (i.t) section acknowledge is working on and hopefully will be fixed soon just letting others know via the forum that these columns are wrong r+s acknowledge it and are currently working to rectify the problem hope this clears it up for you
cheers
garyf

[wesmip1]

I can tell you why its rated second.

Neurals are only as good as the data that is used to train them. In the data they used to train the neural black box a few horses with 0 prize money have probably won. This makes the neural network think that 0 prize money is a good thing. Logic on the other hand tells you this is a bad thing.

I've analysed the neurals from what I believe they actually measure and quite a few of them are not very good. You need to know which ones to use. $$$ has always been a bad rating to use. The others to avoid are DLR, JT, TA and TIM. When I say avoid it depends on which races you use them in.

[garyf]

hi wesmpp1 so another words the ($$$ settings) and the (distance columns )that robert vitlakis the i.t man at r+s has agreed are identifying the wrong read(in those columns only) are actually right according to you so therefore i should ring monday and tell him to leave it alone ??? also how do the neurals identify a horse as having won a race yet recording zero p/money as stated in your last post
cheers
garyf

[wesmip1]

gary,

Its a function of how neurals work. All they can do is retrain the neural network with a new set of training data. Therefore giving different weights to the new pathways. You will find the new ratings can have as many issues as the old ones (might be the same issues might be different issues). It sounds to me the data they input into the neural network was flawed. Its the most common issues with neural networks.

I don't know how many people have looked into neural networks and how they work but its not a bunch of rules to devise the ratings. Its all behind the scenes and is weighted by the training data. The actual weightings on the values passed in are unknown to the users. We just trust the data input was correct and therefore the output should be right.

Just because you don't agree with the patterns it uses doesn't mean its wrong. Logic though tells us it is wrong but then again maybe its smarter then us.

[wesmip1]

also how do the neurals identify a horse as having won a race yet recording zero p/money as stated in your last post
cheer
If i train a network with these 5 records:
Columns are prize money, won

0,1
0,0
0,1
50,0
100,0
33,0
543,0


The neruals was look at the data and say " those with a value of 0 have won twice". Those with above 0 have won nothing. A lower score for Prizemoney must be better.

This is a simple example. More then likely they pass in at least avg prizemoney, total prizemoney, field apm, field total prize money, wins, races, won/lost

I don't know their inputs, but I do know the basics behind how neural networks work.

[garyf]

i actually rang to confirm the data after spending at least half an hour on the phone and going through numerous examples he said to me the $$$ settings and the distance columns were interpreting the wrong data nothing more nothing less he stated that the $$$ column was to identify the 1st down to the last horse in the order of the prizemoney won then too put that figure as a rating so the highest ranked api will always be the horse with the highest rating same as the days criteria horses racing in around the 7 days will always get 20 points down to horses resuming from a spell will get a rating of 1-2 etc no great mystery really maybe you could give him a call to confirm that they are working on these two columns only as again i state they are interpreting the wrong read his words exactly to me on thursday
cheers
garyf

[lomaca]

i actually rang to confirm the data after spending at least half an hour on the phone and going through numerous examples he said to me the $$$ settings and the distance columns were interpreting the wrong data nothing more nothing less he stated that the $$$ column was to identify the 1st down to the last horse in the order of the prizemoney won then too put that figure as a rating so the highest ranked api will always be the horse with the highest rating same as the days criteria horses racing in around the 7 days will always get 20 points down to horses resuming from a spell will get a rating of 1-2 etc no great mystery really maybe you could give him a call to confirm that they are working on these two columns only as again i state they are interpreting the wrong read his words exactly to me on thursday
cheers
garyfgaryf,
this only confirms my reluctance to use ratings I have no control over.

But leave that aside, instead of trying to find complicated explanations, this whole thing could simply be a matter of wrong data feed or even as simple as a "<" instead of ">" in a sorting routine, where the lowest ranked gets to the top instead of the highest.

God knows it happens more often than not, in the world of coding.

Cheers

[garyf]

exactly lomaca my sentiments as well but i believe it is only these two columns but couldn't agree more with your statement what a waste for those who have included these settings recently including yours truly as i don't know when these columns went wrong all the data is irrelevant as wrong points mean wrong order etc wish now i didn't ring the old saying what you don't know won't hurt you should have left well enough alone i think
cheers
garyf

[wesmip1]

i actually rang to confirm the data after spending at least half an hour on the phone and going through numerous examples he said to me the $$$ settings and the distance columns were interpreting the wrong data nothing more nothing less he stated that the $$$ column was to identify the 1st down to the last horse in the order of the prizemoney won then too put that figure as a rating so the highest ranked api will always be the horse with the highest rating same as the days criteria horses racing in around the 7 days will always get 20 points down to horses resuming from a spell will get a rating of 1-2 etc no great mystery really maybe you could give him a call to confirm that they are working on these two columns only as again i state they are interpreting the wrong read his words exactly to me on thursday
cheers
garyf
If that is the case then they are not using a neural network to work out the rating and therefore shouldn't call it as such.

Anyway glad you picked it up.

[garyf]

unsure on that one all that was related to me was the p/money column was calculated on prizemoney won over the last 12 months unlike most services who just give an api for a horses career eg dynamic race-odds column risa i'm led to believe give total p/money won leaving you to divide it by nos of starts for a average didn't discuss the distance column other than them saying the same horse was top rated for the distance column after having one run at 1200 metres ran 2nd last in a maiden now stepped up to a ratings 62 over 1600 didn't ring to query that column they picked it up themselves i'm like you i thought the neurals were based on mathematical computer generated formulas to complicated for anyone to work out even the programmers who knows maybe they are reckon the time algorythm is but after speaking to them i can assure you the p/money isn't at least it's based on last 12 months not career wasted hundreds of hours putting it on a spreadsheet when these two columns are wrong the r+s people are very committed to fixing the problem we are dealing with computers unfortunately am ringing them tomorrow to see the progress am sure the problem will be fixed just not sure how long it could take
cheers
garyf

[garyf]

must remember i didn't advise them till late thursday as everyone knows the data for friday saturday and sunday were already listed on the site so no chance of fixing those 3 days
cheers
garyf

[stugots]

& who said paragraphs & full stops were overrated...

[garyf]

hi all
just a quick post to advise members that i've just got off the phone to R+S they have advised me the( $$$) + the (distance) columns are still giving the wrong read but they are still working to solve the issue but unfortunately with an issue like this could not give me an exact date they advised me to ring back wednesday again which i will do it appears only these two columns are interpreting the wrong data they input for some reason the others columns are working fine
cheers
garyf