When Zero Meant Broken

I remember staring at the scanner button like it had personally insulted me.

MyTradePilot was supposed to look through the market and show possible trading opportunities. That was the whole point. Press the button, get candidates. Maybe good ones, maybe bad ones, but something.

Instead, I got 0.

Not once. Not in a “markets are quiet today” sort of way. Just 0 candidates, over and over, like the app had built a very confident brick wall and parked itself behind it.

At 50, with no coding background, I can usually tell the difference between “I don’t understand this yet” and “this thing is broken.” This felt broken.

So I asked the AI to investigate. It came back with an explanation that sounded very calm and very certain, which somehow made me feel worse.

The problem was in something called mockMarket.ts. There was a broken seeded random-number generator. A hash function was returning a float where an integer was needed.

I’m going to be honest: I understood almost none of that.

“Random-number generator” at least sounded familiar. Like rolling dice inside the app. “Seeded” apparently means it should produce the same kind of fake randomness each time, so the scanner can test against pretend market data in a consistent way. A “hash function” is one of those phrases that makes my eyes slide off the screen. And a “float where an integer was needed” meant… a number with decimals where a whole number was expected, I think.

But none of those words matched how I felt.

How I felt was: I pressed the main button and the app lied to me with a straight face.

That sounds dramatic, but this is a trading tool. The entire thing depends on trust in numbers. If it says there are no candidates, I need to believe that means something. Not that a tiny piece of math underneath got confused and quietly shut the whole thing down.

This was not the only numbers scare. At one point the app showed a $124,000 phantom gain in the P&L. Not real money. Not a real win. Just a terrifying fake number sitting there like it belonged.

That one really got to me.

An app crashing is annoying. A button not working is frustrating. But wrong financial numbers feel different. They feel dangerous.

With the 0 candidates bug, I had to sit in that uncomfortable place where the AI was telling me the fix, but in a language I didn’t speak. It wasn’t being vague. It wasn’t refusing to help. It was just explaining a machine problem in machine words.

And I had to decide whether to trust it.

That is one of the strangest parts of building this way. I am not writing the code by hand. I am directing, checking, questioning, and trying to keep the product honest. But when something breaks deep underneath, I can’t pretend I suddenly know what mockMarket.ts means just because it is on my screen.

So I did the only thing I could do. I asked follow-up questions. I made it explain the bug in plain English. I asked what changed. I tested the scanner again. I checked whether the result made sense.

Eventually the scanner returned real candidates.

Not perfect. Not magic. Just actual candidates, which felt like getting oxygen back in the room.

Later, when the P&L math checked out against a spreadsheet, that was a tiny win I probably appreciated more than I should admit. It wasn’t glamorous. Nobody claps because your numbers add up. But for this project, that mattered more than any logo, screen, or fancy feature.

And yes, I had spent too much time on the logo. I paid for logo generation, rejected one, asked for five more, picked the paper-airplane-and-candlestick idea, then changed my mind and uploaded my own image. Meanwhile the scanner was returning 0 candidates because some random-number thing was broken under the floorboards.

That contrast still makes me laugh a little.

This bug was a wall for me because it exposed the part I’m most afraid of: not knowing what I don’t know. A real user losing real money because of a bug I didn’t catch is a very different fear from “the app looks bad” or “the page is slow.”

I kept going, but I didn’t feel brave while doing it. Mostly I felt out of my depth, asking the same question three different ways until the answer sounded human.

Small takeaway: if the main button returns zero, don’t assume the app is smart — make it prove the numbers are real.