← Back to context

Comment by Esophagus4

3 days ago

A question I have: do you have a background in capital markets?

Because what it looks like is you asked Claude to create a stock picking newsletter based on signals that retail commonly thinks are bleeding edge but institutions have been using for two decades and probably don’t give much edge anymore (AI analysis of news sentiment…)

And you backtested from 2012-2025, one of the longest bull runs in history. Of course your strategy of “buy NVDA” will win in that window.

Which makes me wonder: have you ever worked at a hedge fund?

I have to say I’m tempted to subscribe and have Claude manage an imaginary portfolio based on your recommendations and publish that. Because to be frank, I’m fairly confident your recs won’t generate alpha.

12 comments

Esophagus4

Reply
irldexter  3 days ago

No hedge fund background, fair cop. I'm a software developer who got frustrated with discretionary noise and built something rules-based.

On the bull run, alpha here is measured against SPY, which ran the same period. The claim is relative outperformance, not that we picked a rising tide. The window also includes 2020 and 2022, not great for momentum strategies. Signals that showed near-zero predictive value in statistical validation are zeroed out of the scoring.

Paper trading is already running against live signals. The daily reports are public and free, no subscription needed to check. Methodology is at https://jumpstartsignal.com/how-it-works/ if you want the detail on what's actually being validated.

  • iterateoften  3 days ago

    Unless you know exactly why paper trading sims are so hard to backrest in practice, it’s silly to make arguments on why your paper trading sim works.

    It’s insanely easy to make a trading algo profitable on historical data.

    • irldexter  3 days ago

      Overfitting on historical data is a real risk and defo a concern (there's been lots of learnings lately). The backtest wasn't naive. Fundamentals used filing dates not period-end dates to avoid look-ahead + scoring was validated out-of-sample using walk-forward testing rather than just optimised in-sample (GA used 5 temporal folds and walk-forward used 25 rolling out-of-sample windows).

  • Esophagus4  3 days ago

    One last question - as a percentage of your net worth, how much of your own money have you put into the stock picks generated through this analysis?

    I have a feeling I know the answer, but wanted to check.

    • irldexter  3 days ago

      Well, it's early, so there's only been 2 actionable signals so far; 1 unique ticker SPOTLIGHT and 1 OPPORTUNITY (MONITOR is just a watchlist flag i.e. continue to MONITOR). Am putting in 1K per SPOT/OPP so far. I also maintain ETF and ETC positions (and have good/bad performing stocks since before JSS). So yes, I am putting my money where my mouth is. I have another 10K ready to deploy specifically to this, and other underperforming sources to draw from based on performance.

      1 reply →

  • altmanaltman  3 days ago

    Those emojis on there take away from any seriousness you are trying to project and smells of AI generated text.

    • irldexter  3 days ago

      Hmmm, only did one ascii emoji in a different response. An old habit from the early days. Point noted.