← back to the Zoo
● the receipts ledger · graded dailyloading…
Trained Market Monkey
The Scoreboard
Every pick the monkey makes gets written down with the price that day — then graded later against what the market actually did. No edits, no do-overs. It practices with pretend money until it proves it deserves the real thing.
The Banana Stand 🍌
The fun way to keep score: pretend $100 went into every single call the moment it was made. This ignores trading costs and doesn't compare to the market — it's the "what if I just followed the monkey" number.
🏆 Winners
Finished calls that beat the market (after trading costs), best first — plus the open picks that are up right now. "vs market" means how much better it did than just buying the S&P 500. Click any ticker to see the stock on Google Finance.
🔮 What if the monkey never sold?
Every finished call gets a second, pretend life: we keep watching the stock as if the monkey never exited — no profit target, no stop-loss, just hold to today. Comparing the two tells us whether the exit rules are earning their keep: do stop-losses dodge disasters or sell at the bottom? Do profit targets bank wins or cut runners short?
⚠ Couldn't verify — set aside
Results the monkey refused to trust: the price it recorded didn't match what the stock actually traded that day, a split scrambled the math, or the move was too extreme to believe without a second look. These are excluded from every number above until confirmed — because the whole point is verification, not taking the computer's word for it.
Permission to use real money
The monkey only graduates from practice to real dollars when every box below is checked. This is on purpose — a hot week proves nothing.
The practice account
Imagine the monkey started with a $10,000 practice account and staked 10% of it on every finished call. This is that account's balance after each call closed — the dashed line is the $10,000 it started with. Above the line it's ahead; below, behind.
The playbook — which signals actually work
Every call starts from a signal — something unusual the monkey spotted. As results come in, each signal type gets promoted to golden (proven money-maker), flagged as a trap (proven loser), or stays still testing.
Does confidence mean anything?
When the monkey says "8 out of 10 sure," is it right more often than when it says 6? If these don't line up, the confidence numbers are noise.
🔬 New patterns the monkey is testing
Beyond the playbook above, the monkey mines its own finished calls for new combinations that keep winning (say, "insider buying at tiny healthcare companies"). A combination must prove itself on calls made after it was spotted — until then it's just a hunch and changes nothing. Finding real ones takes a couple of months of calls; that's honesty, not slowness.
Open picks — still playing out
"Bet" says which way the monkey called it: ⬆ rise means it wins if the stock goes up, ⬇ fall means it wins if the stock drops.
Recently finished calls