Backtest Note
How accurate are our predictions? We track every valuation against eventual outcomes and publish the results here.
Status
Early days
Limited Data
We are in Phase 0 with limited historical data. Meaningful backtest results require several years of outcome tracking. We are publishing what we have for transparency, not because it is statistically significant.
Every valuation we produce creates a prediction that can eventually be tested:
- When a company in our coverage has a liquidity event (M&A, IPO, secondary), we compare the outcome to our predicted range.
- We track whether the actual outcome fell within our bull/base/bear range.
- We analyze whether our confidence grades correlate with outcome predictability.
- We publish aggregate results quarterly (once we have sufficient data).
Methodology
How we measure accuracy
Range Containment
Did the actual outcome fall within our predicted range (bear to bull)? We target 70-80% containment rate.
Base Case Accuracy
How close was our base case to the actual outcome? We measure median percentage error.
Confidence Correlation
Do higher confidence grades correlate with better predictions? This validates our confidence grading system.
Directional Accuracy
Were we right about relative valuations? If we said Company A was worth more than Company B, was that correct?
Results
Preliminary data
Results Coming Q4 2026
We will publish our first backtest results once we have sufficient outcome data from our initial engagements.
Commitment
What we promise
- We will publish results even if they are unflattering.
- We will not cherry-pick which outcomes to include.
- We will explain methodology changes and their impact on comparability.
- We will acknowledge when we get things wrong.
- We will continuously improve our methodology based on outcome data.
Want to contribute to our data?
Every engagement adds to our outcome database. Start a valuation with us today.