I found this advice for SDKs, from the Kerwin series, very interesting.
[When reviewing a professional game,] at the end of each engagement, look at the outcome. You know the outcome of the engagement is even. (Even if the division of spoils was uneven enough to decide a game between pros, in an amateur game it can be considered completely even.) Does it look even to you? If not, reconsider your judgment.
I’d never thought of approaching the review of a professional game this way.
Often I’d be thinking of who I thought got the best of an exchange and why, which is fine – but I wouldn’t check myself like this by taking it on faith that the result was essentially even, and questioning why I considered it otherwise,
I still feel a little uneasy with this method, and in fact I was going to post it in the controversial Go opinions thread.