Psychology & process

How does a beginner go from staring blankly at a screen to having a "eureka!" moment like: "I think markets overreact to weekend news"?

You don't discover these things by reading a textbook. You discover them through a deliberate, structured process of Observation. Professional traders develop their hypotheses through three main channels:

1. Screen Time (The Hard Way)

There is no substitute for simply staring at charts. But you don't just stare; you actively ask questions when you see something weird: What time of day did this happen? What day of the week? What happened right before it dropped? Did it drop and stay down, or instantly bounce?

Example: A beginner might notice that every month, right around the time the US Government announces inflation numbers (CPI), Bitcoin goes crazy for 10 minutes, and then often reverses direction.

*Hypothesis:* "If I wait 15 minutes after a major news event, the market will reveal its true direction, and I can trade the reversal of the initial panic."

2. Understanding Market Participants (The Psychological Way)

You are trading against other humans and algorithms. Ask yourself: Where is everyone else feeling pain, and where are they feeling greedy?

Example: Bitcoin has been stuck between $60k and $65k. Thousands of retail traders are buying near $60k and putting their stop-losses just below it at $59,800.

*Hypothesis:* "If the price drops to $59,800, all those stop-losses will trigger. Once those weak hands are flushed out, institutions will buy it cheap. Therefore, I want to place my buy orders exactly where everyone else is placing their stop-losses." (This is known as Liquidity Hunting).

3. Studying Known Anomalies (The Academic Way)

Published anomalies and trader observations are hypothesis sources, not ready-made systems. Their frequency and sign can change across assets, venues, timezones, costs, and regimes.

*Weekend liquidity hypothesis:* Compare weekend and weekday depth, volatility, continuation, and reversal using an explicit timezone and executable prices. Reject the idea if unseen, costed evidence does not survive.

*Session-handoff hypothesis:* Define session boundaries before testing and check whether any conditional reversal remains after multiple comparisons and costs.

*Mean-reversion hypothesis:* Define the reference, distance, entry time, holding period, and risk rule. A large deviation describes the sample; it does not guarantee a snapback.

An anomaly is a measured departure from a defined baseline in a particular sample. It can arise from behavior, market structure, data mining, or chance—and can weaken after publication or a regime change.

The catalog below contains research prompts, not reliable trades. For each one, define the rule, sample, baseline, timezone, costs, multiple-testing adjustment, and locked validation period.

1. The Turn-of-the-Month Effect

Hypothesis: scheduled contributions create measurable flows near recurring calendar dates. Test the stated dates against every other day, define timezone and asset universe, include multiple comparisons, and check whether any effect remains stable out of sample after costs.

2. The 'Buy the Rumor, Sell the News' Listing Effect

Hypothesis: announcement and listing timestamps produce different return and liquidity distributions. Avoid assuming the announcement was tradable before your dataset timestamp, include failed and delisted tokens, and model the severe latency, spread, and access constraints around the event.

3. Funding Rate Mean Reversion

Hypothesis: extreme funding and crowded positioning may change forward returns or liquidation risk. Define 'extreme' without looking at the answer, separate funding carry from directional P&L, and test whether price, OI, basis, and funding jointly add information out of sample.

4. The 'Bart Simpson' Algo Pattern

The nickname describes a shape after it has formed; it is not a causal model. Research it only after defining the impulse, range, reversal, horizon, and eligible universe mechanically, including shapes that never completed.

5. The Asian/US Session Handoff

Crypto trades continuously while participant mix and connected markets vary through the day. Test fixed UTC session definitions and conditional returns without attributing intent. Daylight-saving changes, venue mix, news timing, and regime shifts can manufacture or reverse an apparent handoff effect.

6. The Weekend Fakeout

Weekend depth and participation can differ from weekdays, but neither low volume nor a calendar label makes a breakout false. Measure continuation and reversal under a predefined breakout rule, compare like volatility regimes, and include the cost of thinner liquidity.

7. The 4-Year Halving Cycle

Bitcoin issuance changes on a known protocol schedule, but the market can anticipate known events and each halving has occurred in a different macro, liquidity, and adoption regime. A handful of overlapping cycles cannot establish a guaranteed post-halving path.

Trading loses money through a short list of well-documented biases, and knowing the list is the first (insufficient) defense. Loss aversion: losses hurt ~2× more than equivalent gains feel good — so you cut winners early and let losers run, the exact inverse of positive expectancy. Sunk cost: 'I'm already down, might as well hold' — the market doesn't know your entry. Revenge trading: sizing up after a loss to 'win it back' — emotional doubling at the worst possible moment.

FOMO: buying because price already moved — entering at maximum crowd euphoria, minimum edge. Confirmation bias: reading ten bullish threads about your bag and calling it research. The market pays whoever exploits these in others and starves whoever exhibits them.

A journal turns your trading into data. Each entry records what you saw, what you did, and why — *before and after*. Review it monthly and patterns surface that memory edits out: the setups that actually pay, the hours you trade badly, the rule you break most. Your journal is your only honest counterparty.

A playbook is the forward-looking half: the specific setups you're allowed to trade, written as if instructing someone else — trigger, invalidation, target, size. If a trade doesn't match a playbook page, it isn't a trade; it's entertainment with stakes.

Tilt — trading from emotion after losses (or wins) — announces itself: checking price compulsively, sizing up 'just this once', skipping the journal, deleting the stop 'temporarily'. By the time you notice, judgment is already gone. That's why the protocol is written now, while you're calm, and followed mechanically later.

The standard circuit breakers: daily stop — three losses or −3R in a day ends the session, no exceptions; size-down re-entry — return next session at half size until a green day; cooling period after any rule break — 24h flat, journal review before the next order. The market will still be there. Your capital might not be.