Confluence Case Studies: When Aligned Signals Called the Move, and When They Didn't

How to Read These Case Studies

Each case study describes a historical moment, approximates what the confluence score would have read at the time, and reports what happened. The historical score approximations are backward-looking reconstructions using the same pillar inputs the current engine uses. They represent what the methodology would have output given the data available at the time, not what any real-time system actually reported.

The goal is educational. Seeing confluence setups in context teaches how the methodology interprets multi-pillar alignment better than abstract description.

Case 1: March 2020 COVID Crash Bottom

Context: Bitcoin fell from $9,200 on March 7, 2020 to $3,800 on March 12, 2020 during the global COVID-19 liquidity panic. Equity markets crashed simultaneously. Risk-off cascaded across every asset class.

Pillar readings at the bottom (approximate):

Composite: ~72/100 (Bullish confluence)

What happened: BTC bottomed at $3,800 on March 13 and began an extended rally. By July 2020, BTC was back at $10,000. By late 2020, above $30,000. The confluence setup called the bottom within days of the actual low.

Why it worked: Four of five pillars agreed strongly bullish. The news pillar was the dissenting voice (COVID was dominant negative news), but the macro pillar's read on Fed easing and the on-chain pillar's accumulation signals overrode. The structural forces (Fed liquidity, forced selling exhaustion) that would drive the recovery were visible in the data before the recovery began.

Lesson: When sentiment, on-chain, and technicals agree at extremes, and macro supports them, news headlines can be ignored. Confluence across the deeper structural pillars beats the surface-level noise.

Case 2: November 2021 Cycle Top

Context: Bitcoin hit $69,000 on November 10, 2021, then declined over the next several months. The top was formed quietly with no dramatic single-day collapse.

Pillar readings near the top (approximate):

Composite: ~26/100 (Bearish confluence)

What happened: BTC declined from $69,000 to $35,000 over subsequent months, eventually bottoming at $15,500 in November 2022. The cycle top held. Early 2022 rallies failed to reclaim the previous high.

Why it worked: Four pillars agreed bearish. The news pillar was mixed but not strongly positive. The divergence between sentiment (extreme greed) and on-chain (distribution by smart money) was particularly high-signal. When retail is euphoric and informed holders are distributing, the setup is structurally bearish regardless of how the chart looks short-term.

Lesson: Extreme greed with distribution is a stronger sell signal than bearish technicals alone. On-chain data seeing smart money leave the market while sentiment says "buy" is a textbook top pattern.

Case 3: January 2024 Spot BTC ETF Approval

Context: The SEC approved the first spot Bitcoin ETFs on January 10, 2024. Markets had partially anticipated approval in preceding weeks but with significant uncertainty.

Pillar readings in December 2023 (approximate, one month before approval):

Composite: ~68/100 (Constructive to bullish)

What happened: BTC rallied from $42,000 in December 2023 to $73,000 by March 2024 post-approval. The rally continued through most of 2024. The confluence setup pre-approval correctly anticipated the move.

Why it worked: All five pillars were at least moderately bullish. No pillar was neutral or negative. The composite was high but not extreme (68 vs 90+), reflecting that the rally was priced-in partially. Post-approval, scores pushed toward 80+ as remaining doubts cleared and flows materialized.

Lesson: Mid-range bullish confluence (65-75) often represents the setup phase before a major move. The score doesn't have to be at extreme levels to be actionable. Broad agreement across pillars with room to run is sometimes the better entry than peak alignment.

Case 4: Luna Collapse, May 2022 (Confluence Failure)

Context: Before Luna's collapse, Terra's ecosystem had been growing rapidly. LUNA and UST were among the top crypto assets. Pillars for BTC and ETH in early May 2022 were mixed but not extremely negative.

Pillar readings for BTC just before Luna collapse (approximate):

Composite: ~40/100 (Weak bearish)

What happened: UST began depegging May 7-8, 2022. By May 12, LUNA had gone from $85 to fractions of a cent. The broader crypto market crashed alongside. BTC fell from $32,000 to $26,000 in days. The confluence score did not predict this.

Why it failed: The event that drove the crash (UST's algorithmic stablecoin design failing under stress) was not visible in any of the five pillars' inputs at the time. No pillar was tracking Anchor Protocol deposits, the core economic mechanism that fed UST demand. The sentiment pillar didn't register the building skepticism among DeFi-native traders because it wasn't filtering for that specific concern. The technicals pillar had no visibility into protocol-level risks.

Lesson: Confluence can't see what it doesn't measure. Black swan events originating in protocol-level design flaws, un-anticipated regulatory actions, or specific smart contract failures are blind spots. The score was "weak bearish" before the event and "extreme bearish" after. Before-the-fact, it gave no warning.

Implication: Traders using confluence should always maintain independent awareness of systemic risks outside the methodology's inputs. The score is a read on measurable signals. It cannot substitute for attention to risks that aren't being measured.

Case 5: November 2022 FTX Collapse (Confluence Warning Worked)

Context: FTX was the second-largest crypto exchange by volume. Its collapse in November 2022 was one of the largest failures in crypto history. Unlike Luna, several pillars caught warning signs days before the acute event.

Pillar readings November 5-6, 2022 (approximate, 2 days before CZ's tweet):

Composite: ~36/100 (Weak bearish / growing concern)

What happened: FTX collapsed into bankruptcy November 11-14, 2022. Users with funds on FTX lost access. The broader market crashed, with BTC falling from $21,000 to $15,500. The confluence score didn't predict total collapse, but it was warning bearish days before the event.

Why it partially worked: On-chain and news pillars caught warning signs. Unusual flows from FTX wallets and the CoinDesk balance sheet leak were registering in the scoring. Sentiment was starting to shift. Confluence dropped into the "weak bearish" range, which is a warning to reduce exposure even if it doesn't predict the specific event.

Why it didn't fully work: The methodology doesn't measure exchange-solvency risks directly. Outflow spikes can indicate panic withdrawal OR normal customer activity; context is required. The score said "something is off" without identifying the specific risk. Traders who saw the score decline and reduced positions were protected; those who ignored the warning weren't.

Lesson: Partial confluence signals are still useful. A score dropping from 55 to 35 is a material shift even if it isn't in the "extreme" range. Rate of change matters, not just absolute level.

Patterns Across Case Studies

Looking across the five cases, several patterns emerge:

The clearest signals come from extreme alignment. When four or five pillars agree strongly (above 70 or below 30), the probability of a significant move is high. The March 2020 and November 2021 setups both had four pillars aligned at extremes. Both played out.

Partial signals are warnings, not predictions. When composite drops into weak bearish (30-40) or rises into weak bullish (55-65), it's signaling regime change without necessarily predicting magnitude. Traders should adjust positioning without committing aggressively.

News-pillar divergence matters. When macro, on-chain, sentiment, and technicals agree but news is dissenting (COVID 2020 being a good example), the deeper pillars usually win out over the news narrative within weeks.

Protocol-level risks aren't captured. Luna 2022, certain DeFi exploits, and specific smart-contract failures are not in any of the five pillars' inputs. Independent awareness of these risks is always required.

Scores are for decisions, not answers. Every case study involves a trader deciding what to do with the information. No score executed a trade, managed risk, or took profits. The methodology informs the decision; it doesn't make it.

What the Engine Can't Do

Worth stating directly as part of any case study review:

• The engine cannot predict protocol exploits before they happen.
• The engine cannot predict specific regulatory actions before they're announced.
• The engine cannot predict individual counterparty failures (exchanges, funds).
• The engine cannot predict geopolitical black swans.
• The engine has reduced reliability on thin-data altcoins compared to majors.

Every case of confluence failure in crypto history involves at least one of these blind spots being decisive. Users acting on confluence scores should supplement with direct monitoring of these categories.

Related Confluence Content

For the methodology behind what the case studies describe:

• How the confluence score works: The mechanics of individual pillar scoring and composite calculation.
• Pillar weighting: How pillar weights adjust based on market regime, which influences every case study reading.