This is an idea that keeps rolling around since a few odd posts we have had in the forum and some experimental dog(Greyhound) bets I tried.

First the dog bet stats. tried 28 bets at random (I like this name best!) Result 6 wins 1 non-runner. Its too small a sample to be significant but the strike rate was more than 1 in 6.

Now I saw this system sometime ago for horses which is essence is as follows. I will put the rest of the theory on it in here later.

Monday select and take bet with odds 2/1 Stake 3 units
Tuesday odds 2/1 Stake 3 units
Wednesday odds 4/1 Stake 1.5 units
Thursday odds 4/1 stake 1.5 units
Friday odds 6/1 stake 1 unit
Saturday odds 6/1 stake 1 unit

Total bank divide by 20 gives you new unit.

The idea being that if you have a strike rate more than 1 in 6 bet your betting bank increases over 6 races. If you only have a 1 in 6 strike rate you lose 3points (quoted from memory not checked)

Now I tend to dismiss this as you must get the bet and the odds together. Then I started thinking averages! Suppose you tried to get an average of odds at 2/1 and bet 3 units. Another way of putting this would be look at a dog race and pick the dog with forcast odds of near 2/1. If more than 1 choose the one with the fastest time, if the times are the same then select the dog closest to trap 1. (For example!).

Now forget the days part or the system and work your way in sequence through the days dog races choosing odds as laid out above.

I looked at this on paper last night and got first six no hits on the second six 2 hits.

Any pointers or thoughts about this from dog bettors might be useful or is it a complete system that now needs a testing accross the standard in this case 600 races to see how it works out!