AI & Research Framework · Lesson 01 of 05
The research loop: hypothesis → evidence → invalidation
What it is
Everything before this shelf taught you signals. This shelf teaches you what to do with them — and it starts with the engine that separates research from opinion.
The research loop is the scientific method, pointed at markets:
One — form a hypothesis. A specific, checkable claim. Not "I'm bullish" — that's a mood. Something like: leverage has been flushed, spot demand is holding the floor, so I expect this range to resolve upward within weeks.
Two — gather evidence. For and against. Both readings, always — the habit every lesson in this curriculum drilled.
Three — define the invalidation. The exact condition that would prove you wrong — a level, a behavior, a date — chosen before you commit, while you're still calm.
Four — act or publish. Put the claim somewhere it can't be quietly edited.
Five — let reality grade it. Hit or invalidated, no partial credit for vibes.
Six — update, and loop. Wrong? The market just paid you information. Feed it back and go again.
Now compare that to what most people run: form a feeling → collect only confirming evidence → move the goalposts when it goes wrong → forget the whole thing ever happened. That broken loop is the default setting of the human brain in markets. The research loop is the repair.
Why it matters
One idea does most of the work in this lesson: falsifiability. A claim you can't be wrong about isn't analysis — it's a horoscope. "Bitcoin will go up eventually" cannot lose, which is exactly why it cannot help you. The moment you add what, by when, and what kills it, an opinion becomes a thesis — something reality can grade.
And the invalidation is the crown jewel, for three reasons. It makes you honest: you decided what "wrong" looks like before your ego had a position to defend. It manages risk for free: the invalidation is where a position stops making sense, which is why every recorded setup on this site places its stop there. And it makes learning possible: only graded predictions teach you anything. Ungraded ones just teach you to like yourself.
There's a reason this lesson exists on this particular website. The loop isn't advice RIX Intel gives — it's the machine RIX Intel is. Every Journal publication is one loop iteration made public: the read is the hypothesis, the observations carry evidence links, "what would prove this wrong" is printed in a box, and the Track Record is step five running in public, win or lose. You've been looking at the loop the whole time. This lesson just hands you the blueprint.
The two readings, always taught together
Read bullish when
- For a method lesson, "bullish" means: what a well-built bullish thesis looks like. It's specific. An asset, a direction, a rough horizon, a mechanism. "SOL resolves this compression upward within weeks because spot is leading and funding is flat" — attackable, checkable, gradeable.
- It stacks independent evidence. Structure says one thing, flows agree, positioning agrees — the confluence habit (capstone) as evidence-gathering, with the costliest-to-fake signals weighted heaviest (the on-chain shelf's great filter).
- It names its death. A daily close below the range floor kills this. The strongest bullish sentence in research isn't the target — it's the invalidation, stated plainly, because only a thesis that can die deserves belief.
- It survives its counter-case. You wrote the best bearish argument yourself, and the thesis still stands. (This site won't publish without one. Now you know why.)
Never alone — confirm with the falsifiability card & the counter-case
Read bearish when
- The loop is direction-agnostic — a bearish thesis needs identical discipline, plus one extra warning. Same anatomy. Claim, horizon, mechanism, evidence both ways, invalidation above. "Distribution at the highs" needs the level whose reclaim proves it wrong, or it's just pessimism with vocabulary.
- Beware the perma-bear loop. Goalpost-moving isn't a bull disease. "Crypto is dead" in November 2022 was exactly as unfalsifiable as "we're so early" in November 2021. Doom that specifies no level and no date is a mood, not a position — and moods don't get graded, which is why people keep them for years.
- Respect the asymmetry of shorts. Markets spend more time grinding up than crashing, squeezes are violent (Derivatives 04), and being early on a short feels identical to being wrong. Bearish theses need tighter invalidations and more humility about timing — the loop enforces both.
Never alone — confirm with the falsifiability card & the counter-case
Visual explanation
Real market example
November 2022 — the loop versus the vibe, on ground you know. By now, that bottom is the most-studied coordinate in this curriculum: the accumulation base (Market Structure 04–05), backwardation (Derivatives 06), panic-priced puts (07), the self-custody exodus (Flows 04), the stablecoin drain (05), cooling CPI ahead (Macro 03), the dollar's turn (04), and real yields peaking (05). Eight lenses, one coordinate.
Now watch the two loops process the same moment. The dominant narrative was "crypto is dead." Run it through the card: What exactly? (Undefined.) By when? (Never specified.) What kills it? (Nothing — any rally was "a dead cat bounce," forever.) Unfalsifiable — a mood with a headline. Its holders could never be wrong, which is why many of them were still repeating it at triple the price.
The loop's version of the same moment: "This range is accumulation — supply is being absorbed at maximum fear. Invalidation: sustained acceptance below the November low. If that prints, this was distribution to nobody, and I'm wrong." Specific, evidenced across independent lenses, and killable at a stated level. That thesis could have died — that's the point. It happened, instead, to be right — but the deeper lesson is that only one of these two statements was ever capable of being graded, and the person running it learned something either way. Reconstructing it here isn't a victory lap; it's the loop, demonstrated on the richest evidence set you own.
How RIX Intel uses this signal
This one's short, because the answer is: entirely. The loop is the site's operating system, enforced by structure rather than intention. A publication cannot go out without its invalidation and its counter-case — the checklist physically blocks it. Evidence must link to sources a reader can check. Outcomes are computed against recorded conditions and posted to the Track Record, hits and misses identically typeset. Changes of view are new, linked publications — never silent edits — because goalpost-moving is the broken loop's favorite move, and the architecture exists to make it impossible.
The curriculum you've nearly finished is the loop's input library. The Journal is the loop running in public. The Track Record is its scoreboard. That's the whole product, in one lesson.
Common mistakes
Where this signal ruins people
Holding unfalsifiable theses. No level, no date, no kill condition — a claim that can't lose can't win either. Stamp the card on everything, especially your own deepest convictions. *Especially* those.
Evidence shopping. Collecting confirmation feels like research and is its exact opposite. The counter-case discipline — writing the best argument *against* yourself — is the only reliable antidote, which is why it's mandatory here.
Moving the invalidation. "It's still fine on the monthly" (Market Structure 06's oldest sin). The invalidation chosen at entry, while you were calm, is the referee. Renegotiating with the referee mid-game means you're no longer running a thesis — you're running a hope.
Never logging outcomes. Ungraded predictions leave memory in charge of your calibration, and memory is a flatterer. Write them down, date them, grade them — a private track record, exactly like the public one this site keeps. It will humble you, and then it will improve you. In that order.