Most ranking systems are built backwards. Maybe yours too.
Nobody sets out to do it. Testing on P123 is so fast that every small change pays you back with a better curve: you add a node and the sim improves, you tune the weights and it improves again, and fifty runs later it looks perfect. Then new data arrives and it breaks, because it only ever fit the past.
So how can you tell? Ask your system three questions.
1. Can you explain every node in one sentence? "It tested well" doesn't count. The sentence has to say why the node should keep working.
2. Has your tuning ever seen this data? If you tuned the weights on your whole universe, the backtest already knows the answers. P123's docs give a simple fix for your next round: add the rule EvenID=TRUE while you develop, then switch it to EvenID=FALSE and test on stocks your tuning never touched. (Both halves lived through the same years, so it isn't perfect. But it's a start.)
3. What's left after costs? Run the sim with slippage and commissions on, at the size you'd actually trade. Some ideas only exist in stocks you couldn't buy at that size.
A bad answer doesn't prove your system is broken. It means you don't know yet.
Now the pitch, and yes, it's a paid service. I'm Carlos Morales. Quant Solvings is in the initial lineup of P123's new Investing Coaches program, with 50+ retail and institutional clients since 2025. What I teach, one-on-one, is the opposite order: a hypothesis you can explain first, then the system, step by step, with a reason behind every rule, and only then the test, out of sample, with costs on. You still need the backtest. But it comes last, and its job is to try to break your idea. If you'd rather watch than read, here's the same idea in a 45-second film.
Where you'd start depends on where you are:
- Beginner: over 4,000 factors and no idea where to begin. You finish with a framework deployed in your own account.
- Intermediate: you have a system and want to know how robust it is.
- Advanced: machine learning, with walk-forward validation and stress tests.
- Tailored sessions: one specific problem in something you've already built.
I give every session myself. The notes, the skills you're building and a certificate at the end all stay in one portal. Sessions never expire, and if your first one isn't for you, you can ask for your money back within 14 days, before a second session (conditions). Levels and prices are on the coaching page.
Run the three questions either way. If your system passes all three, you don't need me. If it doesn't, the first step is a free 45-minute intro call, in English or Spanish, to see which level fits, if any.
Book your free 45-minute intro call
Educational content. Not investment advice. Quant Solvings is an independent practice. Services and fees are arranged directly with me.
