Niv Goren

Had the pleasure of being interviewed on the first episode of the Spikeet podcast (also my first ever podcast). Talked about: - How to use AI/Vibe coding to bypass the high barrier to entry and backtest in minutes - Avoiding the overfitting and the Excel trap - When do I shut down a strategy ($20k loss on $SPRB was def a reason) Hope you like it! youtube.com/watch?v=qkttS6…

@paraholik Lol that’s insulting

@hackertrader Nice flip. You are a better discretionary trader than algo trader

Yes $SER algo took a short and I manually covered and switched to long on this private placement + trapping short and holding above vwap

@N0pr3s3n7 Pipe always a red flag for short and when it tried to break below vwap and came back and hold it gave me the confidence that shorts are stuck and it might squeeze

@hackertrader what gave you confidence to overwrite strategy rules ? Was it holding above vwap or the news ?

@ludvigfix @MewAndCrypto @SmallCapSmarts Whats a book I only read MD files

@MewAndCrypto @SmallCapSmarts @hackertrader Read some books and not rely too much on Twitter.

nvm it had lookahead bias. 320 stars on github and it was taking todays s&p 500 constitutes to calculate ratios from 2006. A ton of research time wasted but at least no money lost.

@nigeriangoals I threw the computer to the trash out of anger

@hackertrader Post the updated, correted EC

@MewAndCrypto @SmallCapSmarts Will try to write something

@SmallCapSmarts @hackertrader What roadmap do you suggest a beginner to start a backtesting journey?

@thelaptoplegend Hi David what did i do now

Old me would be edge leaking so hard today but not anymore Not feeding ur ai bots any more yw niv

@SmallCapSmarts To be fair I would fall without vibe coding in this cause I trusted a git

@hackertrader Backtesting is probably by far the worst thing to vibe code

@ian_trades There’s no other way

@hackertrader Appreciate the transparency and updates dude

hmm I'm not sure you should build it... the raw number have 0 predictive power on breakouts success rate, the 200MA have SOME predictive power but far from the numbers you laid out

@SemiCynic 😂

@hackertrader Wait, you mean “make no mistakes” doesn’t cover “do not overfit” 😱

Let the overfitting competition begin

@13_cme The sharpe 2.7 → 21.4 claim looks like a massive overfit lol

@hackertrader Lot of overfitting will occur but there's a small probability of a novel+useful trading strategy with each run.

Lol competing with lies you invented the genre

@quantyboi @pedma7 Of course

@pedma7 @hackertrader I've built something that works OOS and isn't as dumb as you might think, can I dm you about it.

Isn’t it the biggest overfit the world have ever seen?

@kansai_89 @paraholik I honestly don’t think even QM can avoid a pnl chart like that.. @stamatoudism does that backtest looks approximately like your pnl chart?

@hackertrader @paraholik 2. No, but based on that curve I would dig deeper. Not saying it’s trash, but not „good enough“ for going live imo

Always heard about it but was finally able to prove it. Market breadth seem to have an edge trading momentum breakout type strategy (it could be overfit, mistake etc etc). From 48% CAGR and 40% DD to 46% CAGR and 25% DD. Shoutout to the guy behind this: tradermonty.github.io/market-breadth…

@kansai_89 @paraholik 1. Every big gainers have a few adds along the way (pyramiding) 2. Do u know a strategy that is relying on outliers (momentum) that don’t have this flaw?

@hackertrader @paraholik Remove that one biggest gainer per year and the curve will look flat or negative. Imagine not being able to trade that day or having technical issues etc. :/

@paraholik I wouldn’t say one

@hackertrader The strategy depends on one outlier win per year. Difficulty seems significant