Skip to content
Dave Mabe
Close menu
Dave Mabe

Success Story - Angus T.

By Dave Mabe

It's very gratifying to see MabeKit pay off for traders, but every once in a while, a trader will share a story that blows me away.

That's what happened when I received this from Angus T., a futures trader living in Australia. (Used with permission, lightly edited for clarity)

Angus T:

A few notes on what I was able to get from the cruncher

What it helped me do

  • Build a proper feature file. I built entry-time feature files in your column library for all five of my books: 136 columns on NQ and 98 on the others (VIX family, session volume, calendar, VWAP slope, ATR, book state). I'll be using them for years.

  • Turn ideas into real tests. Its candidates became about 60 engine runs across NQ, CL, GC, ES and my RBS book. Each rule was built into the strategy as a real admission gate, with a pass mark set in advance and an out-of-sample check.

  • Find a live edge on NQ. It surfaced a VIX + session-volume rule. Scoped to one cohort (the 7am ET pre-market shorts), it removed losing trades, added about 4% net and cut the drawdown by 21%. It's running on my accounts now as a forward test.

  • Find a sizing lever on ES. When the 20-day VIX change is above +4.5%, ES longs perform better. Adding a second contract on those trades adds about $140k over 12.7 years with no deeper portfolio drawdown. It's heading to forward paper.

  • Rule things out quickly. On CL, GC, and RBS it showed me what doesn't work before any of it went near the live capital. That's valuable too.

The bigger topics it opened up

  1. Filters vs size levers. Most rules picked out weaker trades that still make money. That makes them sizing decisions rather than on/off switches. I now test every gate against a plain size cut.

  2. Scoping beats blanket rules. Rules that fail book-wide can work on a single hour, side cohort, specific market directions.

  3. Era selectors. Year, price level, running counters, and unnormalised ATR can let a rule quietly say "trade recent years", so I now strip those columns first.

  4. Data lineage. The same rule can behave differently on NinjaTrader's VIX and FRED's VIXCLS, so I reconcile the data before trusting a result.

  5. Development never stops. This was the big one for me. Pulling the books apart column by column showed how much was still worth looking at. As long as I'm trading these books, I'll keep developing them.

Overall, it probably gave me more ideas than answers, but I absolutely consider that a good thing. Several of my current studies started from questions the Cruncher raised.

Thanks again Dave!

Dave:

Lots of great info and ideas here!

This is why I laugh when I hear a trader say "oh, strategy X stopped working a long time ago."

Once you dig in, you'll find dozens of ways to improve your strategy.

But the biggest benefit is a steady stream of research ideas for new things to test and new trading edges.

Nice work, Angus, and thanks for sharing with the list!

-Dave

P.S. Are you ready to take your trading to the next level? Learn the right way to create your own trading strategies with MabeKit coaching. Here's what a MabeKit trader I'm coaching recently told me: "I wasn't prepared for this level of consistency in new equity highs in curves from my strategies." - Ryan Get Instant Access