After running some analysis on my ratings for Saturday Metro races over the past 14 months since January 2013 I have found something interesting. Overall my top rated horse has the following results with no additional filters based on NSWTAB dividends:

Races - 2323
Wins - 651
S/R - 28%
Return - 2371
P/L - 48
POT - 2.1%

Now after running through some filters I found that 1 filter had almost identical bets and wins but a very different return. The filter is "Won at the Track", here is what I found:

Not Won At The Track
Races - 1162
Wins - 321
S/R - 28%
Return - 1252
P/L - 90
POT - 8%

Won At The Track
Races - 1161
Wins - 330
S/R - 28%
Return - 1118
P/L - -42
LOT - 4%

With almost identical bets and wins the only conclusion I can draw is that this factor has no actual bearing on whether the horse will or won't win but it has a very big bearing on the price of the horse on the TAB......very interesting.

Yes Mattio there are certainly some profitable niches to be found for those prepared to put in the time looking and having a great searchable database. I found a good one this morning for Back betting.

For a particular State and Favourite rank:

Track condition suitable based on past performance.
77 Bets 23 Wins 30% SR 5.69% POT

Track condition unsuitable based on past performance
13 Bets 4 Wins 30% SR 96.92% POT

You can see here a contrarian view. The bookmakers, analysts and tipsters have studied past performance, seen a favourite running on say a 'Slow' track for the first time, Opened the price higher and the punters have let it drift. Same strike rate, big difference in dividend.

Great fun finding these!

Mattio
Are you able to break your selns down by started ( or not ) at the track?

__________________
Dear Lord
Please let me break even. I need the money.

I was intrigued by your finding and so had a small potter around the database.

WON AT TRACK

(a) the horse isn't a maiden
(b) the horse has raced at the track before
(c) >= 8 starters

Today

................ WAT ............................ NWAT

SR.............10.28% ......................... 7.87%

Div..............$8.10 ........................... $10.37

R.F. ............1.183 .......................... 0.899

Looking at the RF's the WAT has an advantage.

....... moving right along :-

WON AT DISTANCE

(a) the horse isn't a maiden
(b) the horse has raced at the distance before
(c) >= 8 starters

.................WAD .......................... NWAD

SR.............9.7% ......................... 8.25%

Div..............$8.08 ........................... $9.84

R.F. ............1.079 .......................... 0.917

....... and even further .................

WON IN WET

(a) the horse isn't a maiden
(b) the horse has raced in Slow and Heavy before
(c) >= 8 starters

.................. WIW ......................... NWIW

SR.............10.02% ......................... 7.86%

Div..............$7.90 ........................... $10.70

R.F. ............1.240 .......................... 0.910

The sample size is 1,436,167 horses in several races.

It is my opinion that the relationship between Av.Div and Relative Frequency tells the story.