I backtested 2L+ stock signals to build a consensus stock engine. Here’s what I learned.

Hey everyone,

As a retail trader and software engineer, I got tired of the noise in technical analysis. Some people swear by RSI, others rely on EMAs. But when you try to average them, you realise a static average doesn't work because different indicators dominate different market regimes (trending vs cycling).

To solve this, I spent the months building an engine (TickerSignals) that scores 39 indicators across 6 layers daily for 2,500+ stocks. To calibrate the weights, I backtested roughly 2L+ historical signals.

Here are my two biggest takeaways from the data:

1. Standard averages lie: Averaging indicators is useless. When a stock is in a strong trend (high ADX), oscillators like RSI will stay "oversold" or "overbought" for weeks while the price climbs, yielding false signals. Your weights must dynamically shift to trend following indicators (EMAs, SuperTrend) in trending phases and back to oscillators when the market cycles sideways.

2. Transparency builds trust: Most signal services claim they have a "90% win-rate AI algorithm" but hide the math. When I backtested, my high-confidence (>65%) Strong Buy signals hit their target profit levels ~72% of the time. It’s not a magic bullet, but it's realistic. I decided to make the math 100% transparent, you can click any stock on the platform and see the exact breakdown of how all 39 indicators scored and the track record.

I’ve opened up the platform completely free with no paywalls at: tickersignals.info

I’m prepping for a Product Hunt launch soon. If you trade or analyse data, I'd love your thoughts:

• Does a consensus model fit how you analyse stocks?

• What indicators would you add next?

If you'd like to support on launch day, you can follow the upcoming page here at [https://www.producthunt.com/products/ticker-signals/ticker-signals/prelaunch]

Thanks for reading!