Nearly all systems based around form & ratings (if they’re at all different) track conditions, weight, success at track and distance will return a negative POT, usually around the value of the TAB take. Stats are useful, but one needs to find the correct combinations, but then we end up with what everyone else ends up with.



An example of a decent filter, and one which has been discussed to death over decades, and decades is LSW. Approx 15% of all LSW win their next start. There are arguments to suggest that they don’t represent value because they’re overbet ….. no-one can provide proof as to when a horse is overbet. It’s simply a baseless opinion, and might be one of these myths that could actually work in our favour. A LSW might have had bad luck in running, might not have been fit, the jockey might have timed the run perfectly with a minimum of effort required from the horse to win, may not have liked the track or the conditions ….. all the same reasons given why horses get beaten can also be applied to LSW.



The best % place getter in a race is supposedly the only single raw statistic ever found that returns a POT. A POT of 8% apparently ….. see Malcolm Knowles book on Consistency. I don’t doubt the authenticity of the research but looking at the stats used to arrive at this outcome, there’s room for healthy scepticism.



If we combine these two filters we MUST get a profit of greater than 8% !



I’ll betcha ya it returns a negative POT.



Why not a formless system, something “relatively immeasurable”. Many of the decent opinions written here and written in books DO highlight (sometimes obscurely) the importance of the “improver”. Even Malcolm Knowles in his Consistency book “shouts” the improving horse. I wonder how many have “read” this !?



Why not ??

· a certain number of runs from a spell (with or without a win?),

· an increase in distance each time,

· limit the number of starts the horse has had so you don’t have a horse that’s past it’s best,

· maybe 3 y/o (from FEB onwards) to 5 y/o,

· if more than one sel’n take the lightest raced.



Contradicting the above, I do punt one system.

I have tested hundreds of different systems using chromes Race Census program and i can tell you one thing that makes a difference every time that can turn a loser in to a winner.

Price

Give me any single raw filter that you think is a winner and i will test it with 2 price bands, under $3.50 and over $3.50

If you think that the over $3.50 is to general we can put in a max price and i will beat you every time with the under $3.50

Shaun
I have just finished looking at favorites. comparing their SP to their win ratio at that price. I have only just put a line through $3.50 as the point where the Fav becomes overbet. The higher the favorite's odds the more it is overbet. From $2.50 to $3.50 they are only marginally underbet. Under even odds the unders increase to 20% at $1.30. Surprised me. Beton

It can't turn all losing systems in to a winning system but it will improve what ever you have.

Alright here's one. Now this is NO secret at all ..... Malcolm Knowles book Consistency is based on the top place % ranking in any race where the horse has 20 or more starts. He claims an 8% POT.



Gunny72 suggests that the top place % ranking at the moment is a 10% LOT



Maybe you could firstly run the top place % ranking ….. get those results ….. then run the revised results with your price filter.



I don’t know of too many other stats that could be verified that aren’t putting anyone’s system at risk.

The stats i have run from 1999 to 2009 and are for metro tracks only.

Selections 37341
Winners 7601
SR 20.36%
POT -14.61%


<3.50
Selections 11038
Winners 4179
SR 37.86%
POT -9.98%

OK good one Shaun.

How are you finding the database ? I'm seriouslty thinking of investing but cannot find prices or what the parameters are for sorting data ?

Your post on the top % place ranking as a raw stat continues the theme that one raw stat will retun you minus 15% roughly.

I have finally given up on place % as my main system after almost 20 years of using this stat. I have another system that is formless as you say and it has shown good results for just over a year now but not all that much action and you require nerves of steel to follow it. I developed this from trying to modify my place % method and it uses some of the ideas mentioned here by barny. I just dropped the place % angle.

I now use a modified place % method to provide 'entertainment' selections to use in the interim.

For the price he is asking it is a good buy, there are others around that may have more features but are way more expensive, unlike the version i have he now includes country information and you are not forced to update the system every month although new updates are done every three months if you want them for a fee.

He used to have a demo of the program that showed all the info less the data base you could e-mail him here is a link to the site.

http://www.propun.com.au/racecensus_cc.html