Engineering strategy is winning in the long run.
You don't choose the cards. You choose how you play them, and how many hands you get to sit for until you run out of money. Most engineering decisions are made on incomplete information, under time pressure, against a market that isn't even trying to beat you; it's just indifferent. That's a solvable problem, and it gets solved the way poker does: by making decisions with positive expected value often enough that variance stops deciding the outcome.
I write engineering strategy reports and publish them in full, including the parts I got wrong. Every claim carries where it came from. The orange ones are the ones I'm still guessing at.
The framework
An operating contract for engineering strategy: reduce risk, compound the delivery machine, and maximise the probability that the right thing gets built. It's the method behind every report here, free and open.
What it costs to be honest
Every report opens with this meter, before the conclusions. It is counted from the claims in the content, so it cannot drift from what the document actually says.