Hackworth Prime × DOGE

Hackworth Prime is the balanced, all-weather version of AlphaNet's multi-alpha engine: up to 20 low-correlation alpha signals combined and hedged into one strategy, with low average directional exposure and dynamic sizing. “All-weather” is an easy thing to put in a spec sheet. These two months handed it two completely different markets to prove it in.

Test one: the crash

June was brutal. DOGE dropped 28.2% in a month, with multiple sessions down 4% or more. Prime × DOGE returned +10.2%.

This is what the design means by “capturing fat-tails.” Prime doesn't sit at a fixed exposure — it sizes dynamically as its alpha signals shift, and when the market broke hard, the book was positioned to profit from it rather than merely survive it. Week by week, against the asset:

WEEKLY RETURNS, STRATEGY VS. ASSET · HIGHLIGHTED: DOGE'S TWO WORST WEEKS OF THE PERIOD.

Read the whole column, not just the crash rows. The strategy was positive in DOGE's worst week (+4.3% against −18.3%) — and positive again in DOGE's best week (+3.3% against +5.5%). Its worst week in two months was −0.5%. Wins on both sides of the market, losses that stay small: that asymmetry, not any single week, is the result.

Test two: the flatline

Then July took the opposite exam. DOGE spent most of the month drifting in a narrow band — no trend, no momentum, one single air pocket late in the month. Prime returned +5.2% in July so far.

This is the harder test, and the one most strategies fail quietly. A strategy that made +10% in the crash could just be a short bet in disguise — but a short bet earns little when the market flatlines. Prime kept earning because its return doesn't come from one view; it comes from many small, hedged signals firing independently, each incremental trade attributed to a specific alpha. Some of those signals need volatility, some need mean-reversion, none of them need DOGE to pick a direction.

The single best day of the entire period makes the point: +3.28% on 7 July — a quiet day, DOGE down barely 3%, well inside its range. The strategy's best day didn't need a dramatic market. It needed its signals to be right.

One week under the hood

Here's what that looks like at daily resolution — the most recent seven sessions:

DAILY RETURNS, 18–24 JULY · HIGHLIGHTED: THE WEEK'S BEST DAY CAME ON ITS FLATTEST MARKET — AND THE WEEK'S WORST MARKET COST ALMOST NOTHING.

Two rows tell the story. Saturday: DOGE closed four basis points from where it opened — as flat as crypto gets — and it was one of the strategy's best days, +1.22%. Thursday: DOGE gapped down 5.2%, its worst session in weeks, and the strategy gave up just 0.30% — then earned +1.75% on Friday's bounce. That texture — small, frequent, uncorrelated with the asset — is what compounds into +14.2% over two months without the ride making you sick.

The consistency numbers back it up across the full period: profitable on 58% of active sessions, longest losing streak just 3 sessions, and the deepest drawdown of the period — 4.7% — fully recovered within 7 days.

The scoreboard

The drawdown line is the one to sit with: the worst dip in two months was 4.7%, in an asset that fell 31.6% peak-to-trough. Net directional exposure across the Prime book averaged roughly 20% of capital at daily close, which is why the return stream looks nothing like the asset underneath it. And one honest note on the Sharpe of 4.0 — it's real and computed from live daily returns, but 58 sessions is a short sample, and no strategy holds a number like that forever.

The honest comparison

A sharp reader should ask: if DOGE fell 30%, wouldn't simply shorting it have made more? Yes. With perfect hindsight, a static short held from day one would have returned roughly +24% — well above the strategy. We'd rather tell you that than have you compute it yourself.

But that trade only exists in hindsight, and it's fragile in every direction the market didn't go. It eats every rally at full size, and it earns nothing in a market like July's. Prime gave up the hindsight jackpot and bought something else: a return stream that was positive in the crash, in the bounce, and in the flatline, with single-digit drawdowns throughout. You're not paying for a prediction. You're paying for a portfolio of small edges that doesn't need one.

One engine, three temperaments

Prime is one of three versions of the Hackworth engine, each running the same alpha library with a different risk personality:

• Hackworth Prime — the balanced, all-weather version you've just read about: low directional exposure, spread across the full alpha set, built to capture fat-tails in either direction.
• Hackworth Trend — the long-biased, fat-tail-heavy version, built to deliver maximum returns in uptrends while keeping reduced short-side exposure.
• Hackworth OptimaShort — the most risk-averse version, heavier on mean-reversion and shorting, designed for smaller drawdowns at the cost of lower expected returns.

You're not locked into any of them. AlphaNet lets you switch versions freely and start or stop strategies at any time — so the regime call stays in your hands, and the execution stays out of them. If you want the conservative build, OptimaShort ran the same DOGE book with an even smaller footprint. If you think the next leg is up, Trend exists for exactly that view.

See for yourself

Every figure in this piece comes from realised, on-chain P&L — wallet by wallet, day by day, the same data you can open right now on the AlphaNet dashboard. No cherry-picked windows. No simulated fills. And when any strategy hits a rough stretch, you'll see it in the data and read about it here first.

Returns are on deployed USDC, gross of fees and funding, 28 May – 24 July 2026 (58 sessions), computed from daily realised P&L over daily deployed capital and compounded. Sharpe and volatility are annualised from daily returns over a short sample and should be read with caution. Past performance does not predict future results. Digital asset perpetual futures involve substantial risk of loss. Nothing here is investment advice.