It would be interesting to explore the idea of reaction delay as a ranking signal. Perhaps by blind testing it against some of your other established signals of

It would be interesting to explore the idea of reaction delay as a ranking signal. Perhaps by blind testing it against some of your other established signals of quality, and by seeing how well reaction delay can predict these and possibly other ( say, subjective ) signals of quality.