Currently, yes, only brute force.
We’ve looked into other forms of optimization (genetic algo’s/monte carlo/goal seek etc) but the benefits of using these approaches is really dependent on the trading system being modeled and the number of possible variations involved.
Depending on the size of the variation set that’s involved it can be quicker (for example) to do a brute force backtest (than say a genetic algorithm) due to the speed of Seer and the ‘cost’ of the GA. Optimizations under Seer will scale with the number of CPU cores in your box (so almost 4X faster with 4 cores) so the time taken by the brute force approach is less than it used to be.
That said, we’re always open to new use cases.