Does volume really call the bottom?

Burry's June 18 Trading Post — "Dollar Cost Averaging with Volume Signals" — is less about any one stock and more about a method: use share turnover (how many times a stock's float has traded since its peak) to judge whether the shareholder base has rotated into "steadier hands" and the falling knife is near a bottom. His rough guide: 3–5× shares outstanding traded = enough turnover (1.5–3× for old-guard names, up to 10× for newer ones). On that signal he averaged into a basket of wreckage.

The method is clever and it's vintage Burry. It's also got holes worth naming before anyone in here treats turnover as a green light. So — red team.

What he gets right

Capitulation is a real phenomenon, and base-rating shares-traded ÷ shares-outstanding is more disciplined than the usual "feels washed out" gut call. He's explicit that it's a rough supplement to fundamentals, not a standalone trigger. And critically, he's hedged — short PLTR/TSLA and holding puts — so he's not running naked long beta into a knife. The "wait until 20% below my cost to add again" rule also imposes real spacing discipline on averaging down. None of that is dumb.

1. The signal is mechanically guaranteed by the crash itself

A stock down 70–80% will trade multiples of its float almost by construction: panic selling, tax-loss harvesting, margin/forced liquidation, index deletions, and volatility-driven churn all spike volume. High turnover is a symptom of a violent drawdown, not a forecast of recovery. The signal fires loudest exactly when a stock has fallen hardest — which is precisely when it tells you the least about what happens next.

2. "3–5× = bottom" is curve-fit on the survivors

He says he derived the guide by studying past crashes. But the names you study after the fact are the ones that bottomed and came back. Stocks that turned over 5×, 8×, 15× and then kept bleeding — or delisted — don't sit in the sample as clean "it bottomed here" data points. That's survivorship bias baked into the heuristic, and there's no stated false-positive rate. How often does 4× turnover precede a real bottom vs. precede another 50% down? The post doesn't say, because that number is hard and probably unflattering.

3. "Steadier hands" is asserted, not measured

Turnover tells you shares changed hands. It says nothing about who holds them now. Deep falling knives often attract more trader, algo, and gambler flow, not less — the entire "buy the dip on the 80%-off name" genre is short-term money. You cannot infer hand-steadiness from volume. The mechanism he's relying on (trauma-free new holders who won't panic) is an assumption wearing a data costume.

4. It can't tell a rubber band from a marble

This is the big one, and a commenter (KLD811) nailed it under the post: turnover gives every name in the basket the same "confirmed" stamp despite wildly different businesses. A decelerating consumer brand (LULU) is a different species of risk than vertical SaaS (VEEV), payments infra (FISV), or a LatAm marketplace (MELI). Some falling knives are rubber bands that snap back; some are marbles that sink to the floor. The turnover signal is blind to that distinction — and that distinction is the entire game.

5. Six tickers, but not six bets

ADBE, VEEV, and FISV all carry the same bear narrative: "AI / disruption eats the moat." That's one macro bet — "the AI-disruption fear is overdone" — dressed up as a diversified basket. If that thesis is wrong, turnover won't save them; they re-rate down together. Sizing this like six independent positions understates the real correlated risk.

6. LULU's 1,548% cuts both ways

He frames LULU as the "turnover champion" — maximally washed out. But 15× turnover over 2.5 years can equally mean the original growth thesis broke and the entire shareholder base rotated out for good: smart money gone, not steady money arrived. Extreme, prolonged turnover can mark a permanent regime change in the story, not a clean slate. The same number supports the bull and the bear read — which means on its own it supports neither.

7. Averaging down has no circuit breaker

DCA-into-a-falling-knife, then add again 20% lower, is a Martingale-shaped payoff: beautiful when the knife bottoms (Burry's actual career edge) and catastrophic in a structural decline. The only governors here are conviction and a soft turnover read — both subjective. There's no rule in the method that ever says "I was wrong, stop adding." For a guy whose whole brand is being early-and-right, that's fine. For someone copying the trade without his stomach or his hedges, it's how accounts die.

Bottom line — how the gang should actually use this

Treat turnover as a "you're probably not catching the very top of the knife" filter, not a buy trigger. It modestly lowers the odds of being insanely early; it tells you nothing about whether the thing works. The real question — rubber band or marble? — is still unanswered and still requires per-name fundamental work. Respect the correlation: this basket is largely one bet on AI-disruption fear being overdone, so size it as one bet. And remember he's running shorts (PLTR/TSLA) and puts against these longs — copy the longs without the hedges and you're holding a riskier book than he is.

Smart method, real edge in his hands. Just not the green light it can read as. Prod accordingly.