[Verified] +66% a Year — Bull Market Boost or Real Alpha?

The itch to earn more — and the pause that followed

+66% per year, Max Drawdown (MDD) −8.8%

Those are the 5-year backtest numbers for my live active quant portfolio — a 50/50 blend of two systems running in two different markets. The drawdowns are unusually shallow for a return profile like that.

 

And naturally, that’s where the itch starts.

“If drawdowns are this small, isn’t there room to take on a little more risk (more leverage) and push the returns higher?”

I wanted to find the sweet spot of the trade-off — giving up some MDD in exchange for more CAGR.

But cranking up leverage is, in the end, a bet that “this return profile will keep holding.” Before sizing up, I wanted to interrogate two suspicions.

 

---

Suspicion ① — Is this +66% just “the bull market did it”?

2021–2025 was a secular bull run for risk assets. If most of my return came from “the market going up,” then the moment that market rolls over, leverage turns into poison.

Mistaking beta riding a bull market for alpha is one of the quant’s most common forms of self-deception.

 

The test. I regressed the portfolio’s daily returns against a market benchmark and decomposed the return into two pieces.

• Beta: the part earned by tracking the market (and which collapses when the market does)
• Alpha: the part earned regardless of what the market did (real skill)

Results.

The clincher: even in the actual 2022 bear market — the year risk assets got cut in half — this portfolio finished +22.7%.

Across the full five years, there wasn’t a single rolling 12-month window that ended in the red.

Suspicion ① resolved: this wasn’t a bull-market sugar high — it was market-neutral alpha.

Suspicion ② — Then the real risk is alpha decay

Alpha doesn’t die because the market turns. Alpha dies by wearing out.

When money crowds into the same edge, the source of return compresses,

and at some point it just disappears. That’s the quant’s real enemy.

 

I had been believing that my usual approach — dynamic management that periodically rotates what I trade — was offsetting this. But belief isn’t verification…!

My portfolio is built from roughly a dozen independent sleeves, each with its own personality. I compared each sleeve’s risk-adjusted return (Sharpe) for the first half vs. the second half of the sample.

First pass: four sleeves (call them A, B, C, D for short) looked visibly weaker in the second half. “So some of them are wearing out after all,” I started to think.

 

Suspicion ③ (self-check) — Is it really decay, or just “what this market regime is doing”?

The whole point of this post is that I didn’t stop here.------------------aaarghhhhh!!!!!!!!

Doubt, then doubt again..

 

There’s a trap in a naive first-half/second-half comparison.

If an unusually strong early stretch ends up bundled into the “first half,” even a perfectly healthy strategy will look like “its second half got worse.” I had to separate real decay from a recent market regime that simply happened to be unkind to that particular strategy.

 

So I broke the four sleeves out year by year.

A, B, and C were clearly regime effects (market-dependent). All three had their worst year in the 2022 bear market and recovered afterward. The “weakness” wasn’t decay — it was that year’s market environment.

But sleeve D — a sloping-down pattern plus recently negative results — still had suspicion hanging over it. This isn’t a place to wave hands and move on.

 

The final test — skip the polluted metric, look at the signal’s raw edge

An equity-curve-based Sharpe is a polluted metric — leverage, fees, and the market regime are all baked in together. To see whether alpha is still alive, you have to look at the strategy’s raw signal edge itself.

So I re-ran the exact signal logic the bot actually uses, untouched, over the full sample. Strip out leverage and costs, look at the pure edge, and decompose it by market environment (sideways vs. trending).

(Aside: the original reason for suspicion turned out to be a calculation error in the tail segment of a time series stitched together from multiple period-chunks — it had printed a spurious negative. When I reproduced the engine directly, the numbers came out normal. Lesson learned: don’t trust stitched-together summary data; reproduce from source.)

 

Results.

• Consistent every year. 2025 and 2026 are just as healthy as the earlier years.
• Across all four buckets of market environment (sideways / trending) × period (early / recent), every cell is positive. If anything, recent sideways markets have been better than the early ones.

Suspicion ③ resolved: sleeve D was not decaying. The “weakness” I’d seen earlier was just an artifact of stitched-together data.

Conclusion — for now it’s safe, and the watching continues

Of the four sleeves I dug into, zero were confirmed as structural alpha decay. All were either regime effects or simply stable. In other words, the portfolio I’m running today is safe from alpha decay — as of now.

And this isn’t a one-and-done check. A decay monitor runs every day, automatically watching each sleeve’s return distribution for changes. The moment a distribution starts to drift statistically away from the backtest, an alert fires — the point is to catch alpha as it starts to wear out, not after.

Three things to leave you with.

1. Doubt your own returns. Especially when they’re good. The moment you mistake beta for alpha, leverage becomes a weapon pointed at yourself.
2. “Looks weak” and “structurally dead” are not the same thing. If you can’t tell regime from decay, you’ll either throw out perfectly good strategies or cling to dead ones.
3. Verify with the raw signal, not the summary metric. Stitched-together equity curves lie.

 

Only now have I earned the right to return to the original question — “how far should I push leverage to trade CAGR for MDD?” — because I’ve confirmed the returns are real and still alive. The sweet-spot story is for the next post.

---

※ This post is meant to share the system’s reasoning process and verification methodology. Specific signals, tickers, and sizing are not disclosed. Nothing here is investment advice.