There are two definitions of Tufts syndrome.
One is that high stats applicants are rejected in order to protect yield.
Another is that it doesn’t exist (i.e it is a false perception), because high stats applicants are really being rejected for holistic reasons, not yield protection.
https://en.wikipedia.org/wiki/Yield_protection
From a mathematical standpoint, yield is inversely proportional to “the number of students that pass on Harvard” - so making admissions decisions with the goal of reducing “the number of last-minute seats you have to fill” is a textbook definition of yield protection. Note that yield is inversely proportional to acceptance rate, and when your yield is upwards of 82% and your acceptance rate is downwards of 5% even small perturbations in the numbers are important when trying to eliminate the less than half a percent advantage that Stanford has in acceptance rate.
As both you and @skieurope have pointed out, fear of having to fill a few seats from the wait list is not a plausible rational for such behavior, but overtaking Stanford to regain the title of most selective university in the country just might be 
Over enrollment is the result of underestimating yield, so overestimating yield in subsequent years (which results in fewer admits and back filling from the wait list) is prudent and definitely not an example of yield protection.