People forget that stock buybacks also created a pool of equity for employee compensation, it isn't just about returning capital tax effectively.
But basing your expectations on 2021 would have done exactly that. So that argument is wrong.
If your real/amended argument is "no way to predict the level", as in predicting the exact amount, that's a much weaker argument that has barely any ramifications for investing. The specific amount is generally much less important than going over/under some threshold.
But it was obviously a plausible outcome.
Predicting it only took very simple and reasonable logic. "I bet the number stays the same."
A sports analogy: There's a game tomorrow where one team has 30% odds of winning. If someone bets on that team, nobody says "no one could have predicted that". It's obvious that it could happen, and it's plausible that it could happen. Tons of people predicted it.
And large quantities of the stock were purchased. There's tons of trading happening every day.
It's not like the people that expected a large buyback were acting in a vacuum. Stock value follows the average. To push the sports analogy, the team stock would still be going down even though many people are betting on them.
The only surprise with the buyback streak since last year is how poorly executed it was. They dumped $45B ahead of anouncing a decrease in DAU, then trickled after it tanked the stock.
It's cash returned to shareholders.