Behind every whale move is a set of research engines that never stop:
they write new trading ideas, test them on years of real market history,
throw out the ones that don't hold up, and watch the market for the few
that do. This page shows what they did, and what they found.
Loading the engine numbers…
The engines
What survived
Every idea is tuned on 2016–2024 and then judged once on 2025 onward:
market data it never saw. These passed. "Edge" is how much better each
trade did than buying the same stock on a random day.
Idea
Verdict
Edge per trade
Trades
Confidence (t)
Worked on
Working right now
The Recent Edge Scanner re-tests every idea on the last 12 months only.
These are the ideas earning their keep this year, next to how they did over the long run.
Idea
Edge per trade (12 mo)
Trades
Confidence (t)
Worked on
Long-run avg
Prediction Engine
Every weekend the Prediction Engine calls each stock up or down for the week ahead,
using only what was known at Friday's close. When the week is over it checks every
call against what actually happened, and keeps score, including against two
no-skill baselines ("everything goes up" and "same as last week").
This week's calls
Stock
Call
Agreement
Ideas voting
Expected move
Scorecard: right vs wrong
Record
Weeks
Calls
Right
"Always up" would be
"Same as last week"
Right most often, by sector
Sector
Calls
Right
Best and worst predicting ideas
Idea
Votes
Right
Prediction Lab
Where the Prediction Engine gets better. The Lab replays the last three years one week at
a time: to predict any week it may only learn from the weeks before it, using every market
reading Whalur tracks (VIX, crude, gold, bitcoin, the dollar, Treasury rates, the yield
curve) plus each stock's own momentum. Then it moves forward a week and repeats.
