My Calibration

The Calibration page (My Calibration in the nav) is where you study your own track record. The PM Dashboard tells you what you owe; this page tells you how good you are.

It only becomes meaningful once you've resolved a handful of hypotheses. With under 5 resolved you'll see a placeholder asking you to come back later.

Headline stats

Two numbers at the top — the same ones the Head of Product sees when they look at you on the team dashboard.

• Magnitude accuracy — of the claims you got right on direction, what share were also in the correct size band. Direction is the easier prediction; magnitude is the harder one.
• Contested rate — share of resolutions where you ticked I disagree with the mechanical score. A normal rate is a few percent. A high one is a signal that either the scoring is mismatched to your work, or you're systematically dodging accountability.

Blind spots by claim position

This is the most useful chart on the page. It breaks your broke rate down by where in the causal chain the claim sits — what you changed, behavior change, metric movement, business outcome.

The pattern usually slopes from green to amber: PMs are best at predicting their own intervention and worst at predicting the business outcome at the end of the chain. The widening gap is the place to focus.

Read it like this: "I'm 7% wrong on what I changed and 31% wrong on the business outcome — meaning I keep underestimating the steps between my work and revenue. Next time I should add a claim, not a confidence point."

Quarter over quarter

Your held rate by quarter. The question this answers: are you getting better at predicting over time, or stuck in place?

Don't read too much into one quarter — small samples are noisy. Look for the multi-quarter trend.

Filters and view options

The same tag filter bar at the top of the page works here too — slice your calibration to one tag (e.g. only your retention bets) to find where you're strong vs. weak.

The View options button lets you fold private hypotheses into the stats. Off by default; tick it on if you want a fully personal view.

What's not on this page

There's no overall confidence calibration curve here yet — that requires 20–30 resolved hypotheses per confidence bucket, which most individual PMs won't reach. The team-level confidence ribbon does exist on the Head of Product dashboard, where the volume is high enough to be meaningful.

Previous: PM Dashboard ←    Next: The Learning Library →