Is There an AI Stock Picker That Publishes Its Full Track Record, Including the Losses?

July 9, 2026 - 6 min read - TradeWave Research

Most published win rates arrive without the losing picks they were computed from. TradeWave keeps a public daily-pick ledger where all 75 picks since March 17, 2026 sit in the open, the 13 losers marked in red beside the wins. Here is what that page shows and how to check it yourself.

track-recordai-pickstransparencydaily-pickscorecard

By TradeWave Research · July 9, 2026 · AI-Assisted Research: this analysis is generated from TradeWave’s seasonal analytics and reviewed before publication.

Yes. TradeWave publishes one: a daily-pick ledger where every pick is recorded and resolved in public, losers included. Of the 64 resolved daily picks on TradeWave’s public ledger since March 17, 2026, 80% reached the AI’s predicted gain or closed profitable - every pick recorded, losses shown. Held to the window close, 57% were still in profit. Methodology and every individual pick, including losers, are on the public scorecard. Past performance does not guarantee future results.

Most published win rates share a weak point: the losing picks they were computed over are nowhere in sight. The number floats free of its denominator. We built the daily-pick ledger to be the opposite of that - a page where the losers stay pinned beside the winners, in red, with dates - and most of this article is an invitation to go stare at the red rows.

TradeWave public scorecard headline stat block: 80% win rate across 64 resolved daily picks since March 17, 2026, 57% held to close, 4.8% median return. Past performance does not guarantee future results.
The stat block at the top of the live public scorecard, captured July 9, 2026: 80% win rate across 64 resolved picks since March 17, 2026, 57% held to close, 4.8% median return, 75 picks flagged in total. These are the live page's values, not a recreation.

Of the 64 resolved daily picks on TradeWave’s public ledger since March 17, 2026, 80% reached the AI’s predicted gain or closed profitable - every pick recorded, losses shown.

Mar 17 – Jul 9, 2026 · 75 picks flagged, 64 resolved

Past performance does not guarantee future results. Every pick, including the losers.

What Does the Ledger Actually Show?

Seventy-five picks since March 17, 2026. Each was flagged on a dated morning with the AI’s predicted gain attached, and each resolves in public: 64 carry a verdict so far, 11 are still open without one. To be exact about where 64 comes from: the page shows 53 closed rows plus 22 open ones, and 64 resolved is those 53 plus the 11 open picks already marked as having hit their target - a count you get from the rows themselves, walked line by line at the end of this article. The verdict standard is printed on the page - a pick wins if it reached the AI’s predicted gain in the window, or closed profitable - and by that standard the record stands at 80%.

Two more numbers sit in the same stat block, and they are deliberately less flattering. Held to the window close, 57% of closed picks were still in profit, and the median result across winners and losers is 4.8%. All three are shown side by side, so nobody has to take the biggest number on faith. The full walk through what 80% does and does not mean is in how accurate the daily AI pick has been.

Where Are the Losses?

Thirteen of them.

TradeWave public scorecard resolved-picks table with red losing rows visible: 13 of 64 resolved daily picks since March 17, 2026 are marked as losses beside the wins. Past performance does not guarantee future results.
One month of the resolved-picks table on the live scorecard (March 2026), captured July 9, 2026: losing rows sit beside the wins, marked with a red L and kept in date order. Across the full ledger, 13 of the 64 resolved picks are losses.

Scroll the resolved tables and count the red L badges yourself: one in June, four each in May, April, and March. A losing row is formatted exactly like a winning one, sits in the same date order, and never rotates out. No archive page, no footnote, no cleaned-up highlight reel. This figure is the one to slow down on, because it is the part of the page most services do not have.

Why Publish the Losing Picks at All?

Because a win rate is only worth what its denominator is worth. An unblemished record is a marketing decision, not a research result, and the moment losers are allowed to quietly disappear, the number stops measuring the picks and starts measuring the editing.

So the ledger runs on one rule: once a pick is flagged, it stays, and the headline stat is recomputed from whatever the rows say. Today that means 13 losses sitting in the same tables as the wins, and a headline number that changes only when the rows do. The general standard behind this - five checks any service’s record can be graded against - is laid out in how to tell if a track record is real.

How Can You Check the Numbers Yourself?

Do the arithmetic the page invites. The closed tables hold 53 picks: 40 wins, 13 losses. The open table holds 22 more, and 11 of those have already touched their predicted gain - a hit is terminal, so they are booked as resolved wins. That makes 51 winners across 64 resolved picks: 79.69%, which the page rounds to 80. The remaining 11 open picks count for nothing until they close.

That is the whole method, reproducible from the rows in front of you on the public scorecard, no login required. And if you would rather let a script do the counting, the same ledger ships as machine-readable JSON at tradewave.ai/data/daily-pick-ledger.json.

So the answer to the question in the title is one page long. There is an AI stock picker that publishes its full track record, and the proof is not this article - it is the 13 red rows sitting in public beside the wins they keep honest. If you want the daily pick alongside the AI score it comes from, plans are on the pricing page. Past performance does not guarantee future results.

Continue reading