If we could predict the top 5 chances and get odds over the predicted chance then we would show a profit. But its not likely we could do this.

Its an interesting way at looking at it. If you can select 4-5 horses and think they will cover 80% of the market then treat them as one bet and the odds as the dutched prices. So if you could back all of them for $70 then it would be value and worth doing. If you could back them for $90 then you pass on the race. A lot of people do this for various reasons. The best reason being consistentcy. You get lots of small wins and few losses. (100/70 = odds of $1.42).