Place of 49% is a 'working' system standard. If the selections turning up have much less than this it has a good chance of being a looser.

The problem with data banks is they often lack things that would be very important to a Handicapper. 3 'placed' horses can finish with a head or 20 lengths between them and the last runner may only be a length away [or 20] from the winner too. The last 4 starts for example finishing position of a runner is often meaningless.

Running a system through 5 or even 10yrs. of past data and showing a reasonably good win can often be very misleading, as the majority of wins might have been strongly clumped, meaning the system might have only 2 very good winning years and 3 loosing years. Doesn't help you much if the year you start betting is going to be a bad one !!!

Try and have a system tested through 70% of the stats and trailed through the other 30% [far more realistic] and see what happens to the claimed SR and POT. A far more realistic image of the system is then shown rather than just numbers that tell you nothing about how the system will perform in the first 6 months you use it.

My attitude to playing with numbers alone to find selections without any reasoning involved is well known here. I do use systems and make my own but my last rule is always to apply reason and check out the competition. A system can tell you it's selection for a race but tells you nothing about an OBVIOUSLY better horse in the race. Your reasoning can, so most of my serious punting efforts are Handicapping efforts.