We are glad to report that your strategy has been implemented and tested and also very sorry about the late implementation!
Length : Your strategy was implmented with 20 line of code.
Ordering : We ran your strategy on all synthesized designs and ordered them based on predicted scores. The correlation of your strategy’s ordering with the ordering based on the actual scores was -0.125. (1.0 is the best score, -1.0 is the worst score. A completely random prediction would have 0 correlation)
Please note that the numbers specified above will change in future as we’ll rerun your algorithm whenever new synthesis data is available.
More detailed result has been posted on the strategy market page. Thank you for sharing your idea, and we look forward to other brilliant strategies from you!
Hi Jeehyung,
I thought a negative correlation is just as good as a positive correlation because you can just flip the score (100-score) and the correlation would become positive with the same magnitude. A good optimization algorithm should also be able to figure out the correct parameters to automatically achieve this effect of flipping the score. Is that not the case here?
Yes, the negative correlation is just good as the positive.
The reason the optimization didn’t flip the penalty is that, the landscape of the scoring function is very high dimensional and non-predictive (the score is based on sequence patterns!)
I don’t necessarily see this as a bad thing - if the optimization flipped the penalty and we posted it, it would give the false impression that the strategy actually got the positive correlation while it’s actually the opposite of the strategy that gers the positive correlation.
Just to make things clear, I also want to say that the opposite of this particular strategy is NOT reversing the direction of the GC pair - the opposite strategy is to giving bonuses when there are wrong turned GC pairs. Reversing the direction of the GC pair is in fact similar, but different strategy and it’s also in the strategy market page as well.