Systematic strategy research

A backtest is a hypothesis. We track it, live.

Solius searches a large space of systematic strategies on crypto and US equities, then tracks the promoted set live against its own backtest — net of modeled costs, drawdowns and the full roster included.

+20.8%
Flagship strategy · past year
-45%
Bitcoin, same window
+65.7%
Alpha vs Bitcoin
0.76
Sharpe ratio
-13.5%
Max drawdown
3.21
Profit factor

Trailing past year · net of modeled costs · backtest + early live · a research hypothesis under test · figures as of Jul 2026

Flagship equity · past yearBACKTEST + LIVE
Flagship return+20.8%
Sharpe0.76
The method
SystematicRegime-screenedRe-scored dailyMath-only — no AI
01 — The track record

Numbers, not narratives.

Six strategies — three daily risk tiers plus weekly, monthly and quarterly cadences — each built for a different mandate. Over the past year they returned +7% to +21% while Bitcoin did -45%, built on 50 promoted models and re-scored daily. Validated on history back to 2017; every figure is auditable.

Select any strategy for its full track record.

StrategyReturn · past yearAlpha vs buy & holdSharpeMax drawdown
Bitcoin BTC
BTC · Buy & hold
Benchmark
-45%
Flagship strategy · net equity, past year (both rebased to 100)Bitcoin buy & holdBACKTEST → LIVE
Drawdown — underwater · worst -13.5%Bitcoin buy & holdBACKTEST → LIVE
Figures as of Jul 2026Request the full track record

Every figure above covers the past year — return, Sharpe and drawdown alike — so nothing here pairs a recent return with a longer-run ratio. One year is a short sample. The strategies shown are also the top of a large automated search, so selection inflates the record by construction — treat it as a hypothesis, not a promise. Figures combine backtested and early-live periods (live only since promotion), net of modeled trading fees and slippage, with max drawdown shown. Past performance is not indicative of future results. Not investment advice, a recommendation, or an offer of any security or service. Individual results will vary.

02 — The consensus

One consensus per symbol.

Behind the strategies sits an ensemble for each symbol — the equal-weighted consensus of every independently validated model on that symbol, tracked as one signal.

ConsensusReturn · past yearAlpha vs buy & holdSharpeMax drawdown
Figures as of Jul 2026

Each consensus aggregates the full promoted book for its symbol — not a hand-picked winner. Figures cover the past year and combine backtest and early-live periods, net of modeled costs, with max drawdown shown. Past performance is not indicative of future results; not investment advice or an offer.

03 — The engine

Found by machine. Tested across regimes.

Solius searches an enormous space of candidate strategies, then screens the survivors across bull and bear windows — with a Monte Carlo ruin test and a redundancy check against the existing roster. No black box, no AI hype — systematic, mathematical, auditable.

01 — Discover

Discover

Millions of candidate combinations are generated and searched across markets and regimes — far more than any desk could test by hand.

02 — Validate

Validate

Only the survivors advance — a minimum history floor, five validation windows spanning bull and bear, a Monte Carlo ruin test, and a redundancy check against the existing roster.

03 — Rank & retire

Rank & retire

Every strategy is re-scored daily. Conviction rises and falls with the evidence; what stops working is retired without sentiment.

04 — Deliver

Deliver

The output is a daily-refreshed research feed — model scores and risk-tiered strategy allocations, each with a conviction score — over a clean API or dashboard.

04 — Who it's for

Who this is relevant to.

Solius is shared with professionals who do their own diligence — not a self-serve product. It's built for people who read research critically and want a systematic input alongside their own process.

For desks

Funds & prop desks

A daily-refreshed research feed — model scores and risk-tiered strategy allocations, conviction-scored, over API — to run alongside your own process.

For platforms

Platforms & resellers

Research and data infrastructure to embed in your own product or distribute to your book — the engine behind your offering, powered by Solius.

For principals

Family offices & private clients

For sophisticated allocators who want the output of a quant desk without building one — sober, transparent, and reviewed with you directly.

05 — What you get

The output of a quant desk, on tap.

Everything is an output — performance, conviction, delivery. The construction stays under the hood; what reaches you is decision-ready.

Research

Daily model scores & strategies

Risk-tiered strategy research — aggressive to defensive — so you can match the approach to your mandate.

Proof

Transparent track record

A full performance history — backtest and live — you can review line by line.

Delivery

API delivery

Clean, documented endpoints that drop straight into your stack, plus a dashboard for everyone else.

Conviction

Conviction on every score

Each model score carries a conviction weight, re-scored daily — so you can see how much the engine puts behind it.

06 — Why trust it

Engineered to survive contact with reality.

Millions
Combinations tested
an enormous candidate space, searched systematically
Screened
Across bull and bear
five windows, plus a ruin test on every survivor
Daily
Re-scored & retired
conviction tracks the evidence, every day
Math
Only — no AI
no black box, no hype; auditable by design
07 — Questions

Questions allocators ask.

Mostly backtested, and every period is labeled. Each strategy clears a multi-regime screen before it is promoted; forward tracking runs only from promotion onward — recent for all of them — so today the record is overwhelmingly backtest, not a long forward track. Those validation windows sit inside the same history the search ran over, so they are a robustness check, not held-out proof. And the figures are the top of a large automated search, so selection inflates the backtest by construction. Treat the backtest as a hypothesis; the forward track is the test.
Promotion requires clearing a multi-stage screen — a minimum history floor, five validation windows spanning bull and bear, a Monte Carlo ruin test, and a redundancy check against the existing roster — and every strategy is re-scored daily, with automatic demotion on drawdown, alpha, or trade-frequency breach. We do not claim held-out data: those windows come from the same history the search ran over, so no in-sample test can certify forward performance. The only deflation-free check is the forward record, which is why we publish it against the published backtest rather than asking you to trust the backtest.
Everything that keeps them honest: net of modeled trading fees and slippage, max drawdown shown alongside return, and the full promoted roster rather than a hand-picked winner. Figures are re-scored daily and stamped “as of”.
Live results are labeled with their start date and sample size, and the record is moving to a cryptographic, timestamped attestation so it can be independently verified rather than taken on trust. Until that is live, treat live figures as small-sample and early.
Solius is an independent, systematic-research project — not a self-serve product. It is shared with professional allocators and researchers; for questions, reach out directly.
No. Solius is systematic strategy research and analytics — not investment advice, a recommendation, or a solicitation, and not a managed account or an offer of any security or service. You decide what, if anything, to act on. Trading involves risk, including loss of capital.
Contact

Get in touch.

Solius is an independent, systematic research project. For questions or to follow the live-vs-backtest track, reach out directly.