How it works

Disclosed holdings and forecast ranks

Investor pages show public filing history alongside model-ranked expectations for later disclosures. The current Top Picks board uses a shared ranking system; an investor page is not a separately trained model or a claim about that person’s current intentions. This view is identical for every reader. We never execute trades, hold assets, or tailor anything to your account. The tracked roster contains 58 investors.

A US 13F filing is a snapshot of long positions at quarter end, published up to 45 days later. Comparing two snapshots shows the net change in reported holdings, not the trades behind it: a rise from 100,000 to 150,000 shares means the position ended 50,000 larger, while the fund may have traded far more in between, and a position opened and closed inside one quarter never appears. Short positions, written options, futures and swaps are not in 13F at all. Purchased calls and puts appear as separate rows on the underlying security.

Predicting the next buy

13F filings disclose certain positions from a completed reporting period. The current board ranks candidates for a later filing using the promoted ranking model. PriorScore is a within-board ranking score, not a calibrated probability of a purchase or an expected return. The published forecast remains tied to its recorded quarter and as-of date.

Historical tests and their limits

Historical studies use distinct candidate universes, labels and evaluation windows. A walk-forward ranking test measures ordering on its defined holdout; a portfolio simulation measures a different historical quantity. Filing publication delay limits what a portfolio could have known at the time. Methods and caveats are stated with each result rather than generalized across studies.

What the research numbers mean

The historical US research panel reports pair-weighted AUC 0.6762 (95% investor-clustered interval 0.6563 to 0.7002) across 49 managers, 35 quarters, and 124,731 held-out observations. Its top decile was selected 3.17 times as often as the 12.46% base rate. This is a historical ranking result from a separate evaluation. It is not the measured performance of the current Top Picks board, a probability of profit, or evidence of investment alpha. See the sourced methodology and working paper.

A separate historical analysis found pair-weighted AUC 0.696 for the lowest-turnover manager third (17 managers, 290,542 pairs), compared with 0.669 for the highest-turnover third (16 managers, 491,240 pairs). Across 49 managers, turnover and AUC had Spearman rho −0.3388 (p = 0.0173). In the 25-manager subset with at least 200 positive examples, the association was not statistically significant (p = 0.1098). This is a historical subgroup result, not performance of the current board or evidence of investment returns.

Portfolio returns answer a different question. The public consensus backtest reports a small directional average that is not statistically significant. Its result is sensitive to the largest quarter, and the edge does not hold under a realistic filing-availability delay. AUC, historical simulation, and prospective performance are separate evidence classes. This site does not claim that the ranking produces tradable excess returns.

A research publication

Prior Moves is a research publication: one impersonal model portfolio per investor, identical for every reader. You place any trades yourself at your own broker. No execution, no custody, no individualised advice. You hold the trades; Prior Moves publishes the playbook. Nothing here is investment advice. Read the full disclaimer.