Rerun and Stray

Stray works the don't-ignore-it. She pounces on the one number in a neat row that doesn't fit — the surprise everyone else calls "a mistake" and reaches to cross out. "Wait," she says, batting the eraser away. "Why is it different? Maybe it's a slip. But maybe it's the most interesting thing on the whole page — the one place the world did something nobody expected." She circles it. "The results that match what you thought teach you nothing new. It's the surprise that carries every real discovery. So I never throw away the odd one on sight — I get curious about it." She looks over at the woodpecker. "But here's my danger, and I know it. Being curious about a surprise isn't the same as believing it. If I fall in love with every odd result and shout 'DISCOVERY!' the moment I see one — I'll be fooled constantly, because some surprises really are just flukes and slips. Noticing the surprise is only the first half. Testing whether it's real — that part isn't mine. It's Rerun's."

Rerun works the don't-trust-it-yet. She takes the surprising result Stray circled and does the one thing that separates a discovery from a fluke: she does it again. "Stray's right to notice it and right not to erase it," she taps. "But before we believe it — we repeat it. Exactly the same steps. Carefully." She runs it again. Sometimes the surprise comes back — same odd result, second time, third time, and when someone else runs it too. "That's a discovery," she says. "It's real. It repeats." And sometimes it vanishes — the second run is perfectly normal, the surprise gone. "That was a fluke," she says, gently. "A slip, a coincidence, a warm afternoon. Good thing we checked before we believed it." She glances at the cat. "But here's my danger. If I only ever repeat the expected results — if I let someone erase the surprise before I get to test it — then I'll never discover anything. I can only find out whether a surprise is real if someone keeps the surprise long enough for me to repeat it. I need Stray to notice it and protect it from the eraser first."

Then the two of them showed, together, the thing neither could do alone: how a weird, surprising result becomes either a real discovery or an honestly-dismissed fluke — instead of being wrongly thrown away or wrongly believed. A young scientist came, torn. "I got the strangest result," she said. "One plant grew three times faster than all the others. I don't know if I've discovered something amazing or if I just made a mistake. I want to announce it — but I'm scared it's nothing." And it took both of them. Stray spoke first. "Don't erase it — that's the first mistake. You were about to throw it away as 'a weird bad plant.' But why did it grow so fast? That surprise might be a real discovery. Keep it. Get curious." Then Rerun. "And don't announce it — that's the other mistake. One surprising result is a rumor, even a thrilling one. Before you believe it, do it again. Same steps. And let someone else try too." So they did — together. Stray made sure the surprising plant wasn't erased; Rerun ran the whole experiment again, carefully, three times, and had another scientist run it too. And the fast growth... came back. Every time. "Now," they said together, "you have a real discovery — because you did both halves. You noticed the surprise instead of erasing it (Stray), and you repeated it instead of just believing it (Rerun)." The scientist's eyes went wide. "So if I'd only listened to one of you—" "—if you'd only had Stray," Rerun tapped, "you'd have believed your surprise right away — and half the time you'd be announcing flukes." "And if you'd only had Rerun," Stray purred, "you'd have erased your surprise before ever repeating it — and thrown away every real discovery for being strange." "But together," they finished, "we do the exact right thing with a surprise: keep it, and test it. Don't ignore it. Don't trust it. Repeat it — and let the 'again' tell you which it was."