A pat on the back or a kick up the backside

Relating the note from yesterday.

As an office worker, in fact as any worker, you perform tasks every day as requested by managers and clients. Sometimes the same request is directed to you in many different guises. You will do some of those tasks well and some of them poorly.

There are broadly only two influences on the success of those tasks. One is the set of factors that you control and the second is "everything else". You would hope that a manager should only reward or punish for the effect of the first set. But in any one specific task, the infinite possibility of "everything else" means it will probably be the more significant influence. However if given enough time and enough tasks, then your own abilities become statistically valid.

So if you have done exceedingly well in one task, you are likely to also do well in the next one, but statistically not quite as well - even if you are praised. Whereas if you have done exceptionally badly, then, regardless of your punishment, you are still likely to do better next time.

That does not mean a manager has no effect, it means a manager must be aware of his own effect.