Koryu
← Research library
Research · 研究 · 10 · Philosophy

The tide came in. Everyone learned to swim.

13 Jul 20265 min readPhilosophyKoryu Research

The examples below include our own results, deliberately and first. The point of this piece is a habit of attribution, not a claim that we possess skill and others possess luck. Nothing here is investment advice. Decisions are yours.

In a rising crypto market, every strategy works, every feed is brilliant, and every testimonial is sincere. This is the most reliable phenomenon in the asset class, and it carries a brutal corollary: performance achieved during a bull phase, taken alone, contains almost no information about the person or system that achieved it. The tide is not skill. Learning to subtract the tide is.

The decomposition

Any return splits into what the market gave and what the process added. When BTC triples and the median liquid altcoin does more, any long-biased anything prints spectacular numbers, so the honest attribution questions are never "did you make money" but "what did you make beyond survivable exposure to the tide" and "what did you keep when it went out." Skill lives in the residual and in the drawdowns. Headlines live in the tide.

We do it to ourselves first

Our equity sibling published the decomposition of its own flagship backtest, and the numbers deserve to be quoted here because they are exactly the kind marketing usually hides. The five-year result was a 75x multiple; two extraordinary years contributed 18.6x of it, meaning the remaining three and a half years compounded at roughly 4x total, a fraction of the headline rate. Same system, same rules, radically different story depending on the weather. On the crypto side, our founding experiment's regime-gated run more than doubled while its ungated twin lost 22%, and we attribute most of that gated RETURN to the bull regimes the gate happened to keep rather than to any selection brilliance inside them, which is why we present the gate as survival machinery, not alpha, per the transplant write-up. If a publisher will not run this arithmetic on their own results, they are asking you to confuse their beta for their brain.

Why crypto turbocharges the illusion

Three amplifiers make this market the world capital of bull-phase genius. Dispersion: in a strong equity year the index gains twenty-something percent, while in a strong crypto year the median liquid alt can multiply, so even random selection produces life-changing screenshots, and screenshots recruit. Correlation: because everything rises together, any feed of long calls "hits," and inside a single regime a 70% hit rate of tide is statistically indistinguishable from a 70% hit rate of talent. Memory turnover: each cycle imports millions of newcomers who have never seen a winter, a fresh audience for whom every genius is new and every warning is boomer noise. The result is an industry that reconstitutes itself every cycle from the same arithmetic, documented at the category level in the Telegram assessment.

Tool one: benchmark against the regime, not against zero

The question for any crypto performance claim is never "is the number big" but "versus what tide." Two benchmarks do the work: holding BTC over the identical window, and the average return of the universe the trader was allowed to pick from. A 180% year sounds divine until you learn the benchmark did 250% with no fees, no execution risk, and no subscription. Demand the comparison, run it yourself from public prices if refused, and treat a claim that omits its benchmark as a claim that failed it. This is also why our monitor displays market-wide context, breadth, BTC trend state, alongside every asset row: the tide should be visible on the same screen as the boats.

Tool two: demand a winter

A record that spans a full regime change, provably, with the drawdown printed in the summary, is worth more than any bull-phase multiple however large, because winters are where expectancy, sizing, and discipline stop being theoretical. Crypto's winters are not rare tail events; one arrives roughly each cycle, and the 2021 peak to 2022 trough took roughly three quarters off BTC and more off nearly everything else. A provider whose verifiable record starts after the last bottom is asking to be graded on the semester with no exams, which is why the winter question has its own line in seven checks, and why we state plainly that our own young record cannot pass that line yet. Time is the only cure, and we are serving it in public.

The standing mirror

This site will eventually publish strategy results, and this article is the standard they must clear on arrival: regime benchmarks in the headline, decomposition by year, drawdowns stated before returns, and a record long enough to contain weather before it contains adjectives. Until then, the board makes no performance claims at all, and the regime dial exists precisely so readers can do the attribution themselves: when it has read RISK-ON for six months, discount every genius in your feed accordingly, including any you find here. The tide will go out on schedule. Process is what remains. Decisions are yours.

Related reading
PhilosophyThe monitor is not a signal: measurement vs advice, and why the line matters5 min readPhilosophyWhy we sell discipline, not predictions5 min readMethodologyThe transplant experiment: running a stock engine on crypto, in full7 min read
Frequently asked

Why does everyone look skilled in a crypto bull market?

Because dispersion and correlation do the work: when the median liquid altcoin multiplies, any long-biased selection prints spectacular results, and since everything rises together, any feed of long calls hits. The tide is indistinguishable from talent inside a single regime.

How do I separate regime from skill in a track record?

Benchmark against the regime, not zero: compare results to holding BTC and to the average of the eligible universe over the same window. Then demand a record spanning a full regime change, because winters are where process differences become visible.

What is regime attribution?

Decomposing a return into what the market gave and what the process added. A 180% year during a phase when the benchmark did 250% is negative value added; the residual beyond survivable market exposure, and behavior in drawdowns, is where skill lives.

Do you apply this standard to your own results?

Yes, first. Our equity sibling published the year-by-year decomposition showing its headline multiple was dominated by two exceptional years, and our crypto experiment's gated gains are attributed mostly to regime capture rather than selection skill. The standard exists to be applied inward.

What should I do during a bull market then?

Nothing here is advice, but the analytical habit is: discount every genius by the regime, including yourself; write down what part of your results a benchmark would have delivered; and decide sizing rules before the season changes rather than after.