I've had a look at the data for the last 2 years, it proved a bit trickier than I inittially envisaged.
As you suggested, ignoring those at the top of the market will improve returns.
Ignoring those horses with forecast odds of 7/2 and below gives an AE of 1.05 with 1,117 from 9,558 runs with a profit of 296 points. Not a massive return but a good starting point.