As it stands it requires live data to see if an idea or concept has 'legs'

I've searched for horse racing backtesting software and there was one I found, but it's just not BetAngel, so it's not as good in my opinion

The possibility of running some sort of conceptual automation logic against a huge dataset and acquiring performance stats by race type, class and distance would be the holy grail and perhaps worth investigating. If another company can obtain the rights to historic data from BetFair then it should be feasable. Even a database of BSP/ISP/Hi/Lo that could be logically tested against an automation concept would be useful if the database overhead was too great for the whole day.
I'm quite happy to be put in my place and informed that it's a bonkers idea, but, as my paper trade testing takes an awfully long time, I'd appreciate it if someone could point me in a direction where the timescales could be reduced.