Nope much more insidious. A/B testing is statistically random. This thing dangled the desirable result down a path of rewards to attempt to reprogram people into accepting version A while discarding version B, then measuring which method to do so was more effective.