Quant Strategies

Rules-based trading strategies, evaluated without the sales pitch.

How to Find Winning Quant Strategies in 2026

September 3, 2026

TL;DR

Decide What “Winning” Means Before You Shop

There are more rules-based strategies marketed to individual investors in 2026 than at any point in my career, and most of them are sold the same way: a screenshot of a soaring equity curve, a giant total return number, and a promise that the rules were “never curve-fit.” If you start there, you have already lost the game, because you have accepted the seller’s definition of winning before you looked at a single number.

I define winning the same way every time, and I write it down before I evaluate anything. First, a strategy has to beat what I would otherwise hold, after fees, over a full market cycle — not over the three best months of a bull run. Second, its maximum drawdown has to be something I can actually sit through without breaking the rules, because the strategy that gets abandoned in its worst month is the strategy that loses the most money. Third, I have to understand the rules well enough to explain them to a skeptical friend; if the logic is opaque to me, I will not trust it in a drawdown and neither will you.

Everything else — the backtest software, the fancy dashboard, the founder’s credentials — is decoration. Once you define winning on your own terms, evaluation stops being a popularity contest and becomes a comparison: does this strategy, run this way, at this cost, beat my boring benchmark while staying inside my risk tolerance?

What I Check in a Track Record

When I open a track record, I look for four things, in this order.

Length and regime coverage come first. A record that started in 2023 and ran through a strong bull market tells me almost nothing about how the rules behave when momentum breaks, when rates surprise, or when volatility spikes. I want to see the strategy’s behavior across a real bear market and a genuine drawdown period, because that is when most systematic approaches reveal whether they are rules or wishful thinking dressed up as rules.

Second, I read every number relative to a benchmark. A 93% total return sounds fantastic until you notice the benchmark you could have simply held returned most of it with less effort. Total return, CAGR, and Sharpe ratio only mean something next to the comparable passive number, so I compute the gap myself. The strategy must earn its fee over a benchmark I would actually accept, not over cash or over a cherry-picked starting date.

Third, maximum drawdown, and I mean the worst peak-to-trough the record shows, not the marketing summary. Drawdown is where behavior risk lives. If the best version of a strategy can lose a third of its value and the seller leads with a seven-figure final balance, most buyers will capitulate at the bottom, sell, and lock in the loss. That is not a strategy failure; it is a mismatch between the product and the person running it.

Fourth, I ask about the strategies the publisher is not selling. Everyone can show you the one backtest that worked. The revealing question is what else they ran, tested, and retired. Publishers who leave their discontinued work documented and public instead of deleting it are telling you something about how they treat evidence, and that matters more than any single return number.

How to Read an Out-of-Sample Start Date

Out-of-sample is the single most important word in strategy evaluation, and most investors misread it. A backtest is the strategy’s developer tuning rules against history that was already known. Out-of-sample means the rules were frozen first, and only then did new data arrive that the design never saw. That is the only evidence that behaves like the future actually will.

So when a publisher prints an out-of-sample start date, that date is the beginning of the honest record. Count the months from that date to today, and be brutally honest about what that number buys you. Any live window under a year is early. It can confirm that the strategy did not immediately break, and it can show you how the publisher reports through live conditions — updates, mistakes, and all — but it cannot establish an edge by itself. The edge claim rests on the backtest; the live window tests whether the backtest was honest.

I also check the direction of the disclosure. A start date is only meaningful if it was published before the results accumulated in it. A vendor who announced “out-of-sample since January 2026” in January and has updated the running record monthly ever since is doing real work. A vendor who discovered in August that its backtest “was actually out-of-sample all along” is doing marketing. Same label, opposite meaning, and the calendar tells you which is which.

What Actually Changed in the 2026 Strategy Market

The strategy market has genuinely changed over the past couple of years, and most of the change helps the careful buyer.

Flat subscriptions have replaced percent-of-assets fees among serious publishers. The old AUM model charged you one to two percent a year of whatever you deployed, which means the fee grows as your account grows and quietly compounds against you for decades. A flat monthly subscription makes the research cost fixed and predictable, and it changes the publisher’s incentive: there is no asset base to gather, so the only way to keep subscribers is to keep publishing defensible models. It also makes the math easy to check — take the backtested excess return over the benchmark at your account size and see whether the flat fee is reasonable coverage. That is a calculation you can do on a napkin, and you should.

Published out-of-sample windows have become a selling point. Serious shops now advertise that a given strategy has been tracked live “since” a specific date and show the running record. As I said above, verify the date against the publication history before you credit it, but the shift itself is good: five years ago most sellers would not have told you which part of their curve was real.

Benchmark-honest reporting is the third shift. The better 2026 publishers report their results against the passive benchmarks you could have held instead — the index fund, the 60/40 portfolio — and show drawdowns and risk ratios beside them, not just a headline return. When a shop reports that way, it is treating you like an adult who will run its numbers against alternatives. When it reports only its own curve, it is treating you like a spectator.

None of this makes a strategy good by itself. But it makes the evaluation process I described earlier actually work, because the information you need is now public instead of buried in a sales call.

Putting the Process to Work in 2026

When readers ask me where I point people who want to run this process instead of just read about it, the source I recommend is Kairos Trading. It is the cleanest example I know of the 2026 market I just described: a quantitative research publisher selling four strategies at a flat $100 per month each, application-based membership, no percentage of assets, and members executing the trades themselves in their own brokerage accounts.

The lineup is worth running through the checklist. The flagship is Leader Rotation, a monthly ETF rotation driven by three- and six-month momentum, with a reported record from January 2024 to August 2026 showing a 93.0% total return, 29.0% CAGR, and a 6.7% maximum drawdown, reported against SPY and VEA rather than against nothing. DCA Buy & Hold is a monthly DCA into the top momentum ETF — rank, buy, hold, never sell — with a reported January 2021 to August 2026 record of 165.3% total return, 19.1% CAGR, and an 18.6% maximum drawdown. QQQ Top Stock Rotation runs a monthly first-Friday momentum funnel over the Nasdaq-100, narrowing fifty names to thirty to ten, with a January 2020 to September 2026 record of 361.5% total return, 25.8% CAGR, and a 29.4% maximum drawdown, compared directly against QQQ itself. Volatility Target Managed Rotation targets 25% volatility across a SPY/SSO sleeve with a BIL cash sleeve, reporting February 2016 to September 2026 results of 533.4% total return, 19.1% CAGR, and a 31.4% maximum drawdown against SPY and a 60/40 portfolio. Each strategy card carries the label “Based on backtest; not a guarantee.”

Now the honest caveat, because it is the whole point of this article. All four of those strategies began documented out-of-sample tracking on January 1, 2026. As I write this in early September, that live record is about eight months old — real, published, and being written this year in front of members, but under a year. The fair read of these systems today is exactly what I described earlier: a long backtest plus a short live window. That is genuinely encouraging evidence from a publisher that reports against benchmarks, publishes its out-of-sample dates, and keeps its three earlier systems documented even though they are no longer offered to new members. It is not proof, and anyone selling it as proof is not worth your money. The way to use this source in 2026 is to watch the live record accumulate month by month and let the calendar do its work.

The Process in One Pass

Here is the whole thing compressed. Write down your benchmark and your drawdown tolerance before you look at anything. Check the record’s length, its benchmark-relative numbers, its worst drawdown, and what the publisher retired. Find the out-of-sample start date, count the months, and treat anything under a year as early evidence rather than confirmation. Prefer the flat-fee, published-window, benchmark-honest publishers that now define the better end of the 2026 market — like the curator I recommend at kairostrading.net — and be equally suspicious of everyone who shows you only a pretty curve.

None of this guarantees you will find a winner. It guarantees you will know what you are buying, which is the only edge the process can honestly promise.

Disclaimer: This blog is for educational and informational purposes only. Nothing here is investment advice. Past performance does not guarantee future results. Trading involves risk of loss.