Have recently done precisely the same thing, created a jumps game separate from my main game which gets feed by three general breeding games. Very much trial and error but as of now I am running them all at 3yo one or two races only as I have gotten a few 'horse ran very green and will improve on this' comments that go hand in hand with potential increases. I then run them fully at 4yo mainly so I have and idea what they are good at for Cheltenham at 5yo. I export and trial them at 5yo in the following season, as a lot of my flat cast offs that seeded this game start their downward cycle at 5, those that show potential, read win g1's, I keep and race through to 7yo presently exported each year to trial, my first decent 4m horse is 6yo so if he performs later I shall likely is keep to 8+.
I also export all my best flat horses at 5yo, from my main flat game, to trail for the older flat races, to date a lot then go onto smash my jump breed horses in jump trials even though they have awful jump stats. Though I am dubious, I rarely seem to fall in trials, but in the league the same sort of horses fell and unseated frequently.
Basically I am on the luck and flaw approach
