September 26. BTC prints $63,800 on my screen. Spot has ground higher for eleven consecutive sessions, and the Binance perpetual funding rate sits at 0.004% โ flat, occasionally flipping negative for a few hours at a time.
That number is wrong. Every macro leg up since 2020 had funding pinned at 0.05%โ0.1% by the equivalent point in the move. Retail leverage was the tell. Now the tell is silent.
I flagged it as a data anomaly, not a buy signal. Two days earlier, a former FTX community partner named Benson Sun had published the same observation โ wrapped inside a self-built indicator he calls ILI, the Institutional Liquidity Index. His read: this is a "gradual bull market." Yellow divergence means trim. Red divergence means stop trading entirely.
After forty-eight hours of replication work, here is what I have. The funding observation is correct. The framework built on top of it has no audit trail. And its author has an undisclosed book.
The anchor dropped, but I was already airborne.
Context: what ILI actually claims to be
Sun is not a nobody. He ran community at FTX before November 2022, which gives him both a distribution advantage and a credibility discount. He now publishes market commentary, and ILI is his flagship construction. The pitch is clean: traditional cycle-top indicators โ MVRV Z-Score, funding rate extremes, futures basis โ have gone blind, because the marginal buyer changed. Institutions do not FOMO. They buy spot slowly, and the sophisticated ones run Delta Neutral books: long spot, short perp, harvest basis. That structure mechanically suppresses funding and premium, which means "high funding = top" is dead as a heuristic.
His replacement is a composite of three inputs: dollar liquidity, the mNAV premium on corporate treasury vehicles like MSTR, and spot ETF net flows. Blended into one number with three output states. Synchronized: healthy. Yellow divergence: BTC makes a rolling 30-day high while ILI fails to confirm. Red divergence: ILI breaks below 50.
I have built trading systems for nine years. I respect composite indicators. I do not respect indicators I cannot rebuild.
Core: the replication attempt
I pulled the three components into a notebook. Dollar liquidity is public โ Fed balance sheet, reverse repo, TGA. ETF flows are public, daily, from the issuers. mNAV requires market cap divided by treasury NAV, which is also public if you are willing to do the work.
So I built it. Roughly forty lines:
ili = (w1 * norm(dollar_liq) +
w2 * norm(mnav_premium) +
w3 * norm(etf_net_flow))And it died there. w1, w2, w3 are not published. The normalization window is not published. The threshold of 50 has no stated denominator โ 50 of what? A z-score centered on a two-year lookback? A percentile rank against a five-year distribution? Sun does not say, and the original piece does not say either.
An indicator whose weights are private cannot be independently verified. It can only be believed. That is not a technical objection, it is a category error. Chain analytics give us MVRV, SOPR, realized cap, exchange netflow, stablecoin supply. All reproducible. All falsifiable. You can argue with them. You cannot argue with a black box.
Then I looked at the yellow divergence rule and the problem got worse.
Yellow divergence fires constantly. The condition is "BTC makes a rolling 30-day high while ILI does not confirm." In an uptrend, BTC prints rolling 30-day highs repeatedly โ sometimes six or eight times in a single quarter. Liquidity composites are slow-moving; dollar liquidity and Treasury NAV do not reprice on a weekly candle. So the composite will lag spot most of the time by construction. The signal is not predictive. It is arithmetic.
Do the math on signal frequency. If yellow divergence triggers on maybe 60% of new highs, and red divergence requires a full ILI break below an unpublished floor, then you are left with an indicator that is almost never wrong and almost never useful. It is unfalsifiable dressed as rigor. This is hindsight-fit risk, and I have seen it kill more funds than bad market calls ever did.
Here is the part that survives contact though, and it is the part people are skipping past.
The Delta Neutral suppression thesis is structurally real.
I have traded basis. If a meaningful share of new institutional flow arrives as spot-long/perp-short, it creates persistent selling pressure on the perp leg whenever funding goes positive. Funding gets pinned near zero. Basis compresses. The classic "everyone is long and euphoric" readout vanishes because the euphoric longs and the hedged shorts now occupy the same P&L. In 2021 I watched funding hit 0.3% daily at the top and knew. In 2024 the top, if it came, would arrive with funding at 0.01%.
That is a genuine structural change in price discovery, and Sun deserves credit for naming it. My own momentum book flagged the same anomaly in August โ the model that had a Sharpe of 2.1 backtested over five years was underperforming precisely because one of its inputs had structurally decayed.
Where I part ways is the mNAV flywheel.
Sun treats the mNAV premium as an input to a bullish composite. I treat it as a reflexivity bomb. MicroStrategy-style treasuries raise capital at a premium to their holdings, buy more BTC, and the premium widens โ a flywheel with a funding market attached. It works until the equity window closes. Then the marginal bid disappears, and the same vehicle becomes forced supply. That is not a top indicator. That is a liquidation cascade waiting for a capital markets freeze.
A composite index that includes mNAV as a positive input will read "healthy" right up to the moment the flywheel inverts. It has the sign wrong at the terminal node.
Then there is the siphon. Institutional vehicles buy BTC and nothing else. Spot ETFs do not touch DeFi. Corporate treasuries do not touch altcoins. Every dollar that enters through a compliant wrapper exits the alt complex. From 2020 to 2021 I ran mempool arbitrage during DeFi Summer โ $45k of flash loans, $12k net profit in under three minutes off an oracle-timing gap. That market structure existed because capital was fluid across the curve. Today it is not fluid. It is channelized. The BTC dominance chart is the receipt.
Contrarian: the title and the position disagree
Here is what nobody is saying out loud. Sun's headline says gradual bull market, new highs, stay long. His disclosed operations say raise BTC allocation, cut altcoins, trim on every yellow divergence.
That is not a bull. That is a distribution schedule with a bull-market title.
I have been on both sides of that gap. When a trader's narrative and their actions diverge, believe the actions. Every time. The title buys reach. The trim list is what he actually does with his own capital.
And there is a broader tell. "This time is different" is the single most expensive sentence in markets, and "the old top indicators don't work anymore" is its 2024 translation. Institutional flow changed the rhythm of the cycle. It did not change the leverage. mNAV flywheels are leverage. Delta Neutral books are leverage. Basis trades are leverage. The only thing institutions removed was retail's FOMO, which means the top will be slower, quieter, and much harder to see โ not absent.
Consensus is not a metric. But when "institutionalization" becomes the default explanation for every price level, that is a sentiment reading all by itself.
Takeaway
Watch three things, and none of them require trusting a private index. ETF net flow โ if it prints negative for two consecutive weeks, the institutional bid is dead and the narrative collapses with it. MSTR mNAV โ if the premium compresses toward 1.0, the treasury flywheel is unwinding and forced supply follows. Funding rate โ if it stays pinned near zero through new highs, the Delta Neutral thesis is confirmed, and you should stop reading sentiment from derivatives entirely.
Build your own composite. Publish your weights. Then argue with results.
Chaos is just a pattern waiting for a faster eye.
