Users told me my product's core feature was double-counting itself. Here's how I regrouped it, and I

Building a crypto signal bot. The core feature scores how much a set of technical indicators agree on a given trade, higher agreement, higher confidence. I posted it in some trading communities and got feedback that stung because it was right. The problem: most of those indicators are calculated from the same underlying price data. RSI, MACD, EMA, they're different math on the same input, so they largely move together. Ten of them "agreeing" wasn't ten independent confirmations