Bitcoin’s tape over the past 24 hours looked engineered for crypto investors, as BTC surpassed the $90,000 threshold in the early hours of Dec. 29, only to give back those gains less than 12 hours later.
Traders like TedPillows posted clown emojis alongside charts showing repeated peaks and troughs, while CryptoSeth called it “fraud commodity” behavior, pointing to the same sawtooth pattern replaying 30 times.
Additionally, Wimar X blamed Binance and Wintermute outright, claiming “multi-billion dollar manipulation” visible on-chain. However, the on-chain transfers involving Wintermute, as shown in his screenshot, totaled less than $30 million.
Still, the question isn’t whether the accusations are baseless, but whether the data can distinguish between opportunistic stop-hunting and a structurally fragile, overleveraged market that breaks the same way every time someone leans on it.
The microstructure tells the story
Binance’s cumulative volume delta, which is buy-aggressor volume minus sell-aggressor volume accumulated over time, shows a clean pattern: sharp intraday spike driven by aggressive buying, CVD surging as market orders lift offers, followed by an equally sharp reversal driven by aggressive selling, CVD collapsing as traders hit bids.
Price ends roughly where it started, net CVD close to flat over the full window.
That is exactly what a “push through the book, harvest stops and late momentum, then fade it back” sequence looks like. It’s not a slow trend-building conviction, it’s a fast up-and-down that leaves the market roughly unchanged but would be profitable for anyone who traded both legs.
The tape doesn’t show who initiated the move or whether it was coordinated, but it shows the move itself was driven by aggressive directional flow, not passive order matching. These are indicators of market manipulation.
This isn’t a one-off print. The same V-shaped spikes and retraces played out across Bitstamp and Bybit through December. Different venues, similar pattern, repeated over time.
That suggests the environment itself is friendly to exactly the behavior traders are accusing: a structurally fragile, overleveraged market where someone keeps leaning into obvious stop zones because it keeps working.
It doesn’t prove the same trader each time. The market is easy to push around for anyone with enough size and speed to move price in a thin book, then rebalance inventory and collateral across venues before the move reverses.
Someone is stop-hunting
The tape strongly resembles a classic stop-hunt, as liquidity is thin during the holiday period. CoinGecko data shows that Binance is consistently staying below $10 billion, while other major exchanges have even failed to post $1 billion in volume recently.
Additionally, Coinglass data shows that open interest changed by 0.08%, -0.67%, and 0.03% in the past 1 hour, 4 hours, and 24 hours, respectively.
Liquidations over those horizons totaled tens of millions of dollars, split between longs and shorts, not the enormous one-sided wipeouts that accompany a massively crowded trade getting detonated.
Prices at other venues broadly tracked Binance rather than disconnecting, indicating the move wasn’t isolated to one order book. And the on-chain snapshots show custody reshuffling, not the side of the trades or the profit-and-loss path of any particular wallet.
Professional desks were active, as on-chain data shows over 87 BTC exiting Binance to a Wintermute deposit wallet, but what they were doing and why remains opaque.
Taken together, the evidence fits the pattern of opportunistic profit-seeking in thin order books. Aggressive buying drives Bitcoin into a sharp intraday spike, aggressive selling walks it back, and cumulative flow ends up roughly flat.
Repeated inverted V-shaped moves across Bitstamp, Bybit, and Binance, plus a burst of cross-venue flows from Binance to market-maker and exchange addresses, all point to a market that’s easy for well-capitalized traders to push around for short-term profit.
The evidence suggests opportunistic manipulation of the tape. The behavior traders describe is plausible and supported by the pattern, but the data doesn’t identify a specific orchestrator or show intent beyond a reasonable doubt.
What the data does show is that the environment is structurally vulnerable to exactly the kind of stop-hunting traders are accusing, and that the tape looks like someone took advantage of it.