Secrets to Short-Term Trading Ch. 2: Reading the COT Report Without Overreading It

Investment Background

Larry Williams was one of the earliest and most vocal retail advocates for the CFTC's Commitments of Traders report — free, public, and updated weekly, breaking down who holds what positions in futures markets across commercial hedgers, large speculators, and small traders. His central claim, repeated across decades of writing, was straightforward: commercial hedgers, the producers and users of the underlying commodity, are the market's best-informed participants, and tracking their positioning at statistical extremes is a genuine edge.

That claim is directionally sound and remains a confirmed, measured gap in this library — nowhere else does this platform explain COT data honestly. But the honest explanation has to include the part promotional trading material usually omits: commercial positioning is a slow, patient signal, and reading it as a short-term timing tool is the single most common way retail traders misuse a genuinely good dataset.

The Wall Street Translation

What the Report Actually Shows

The COT report classifies open futures positions into three broad groups weekly: commercial hedgers (companies with a real business need to hedge — a wheat farmer, an airline, a gold miner), large speculators (funds and managed money with no underlying business exposure), and small speculators (the retail residual).

Commercials tend to be positioned correctly at extremes over a multi-month horizon — when commercial hedgers are net long a commodity at a level rarely seen in the report's history, prices have historically been more likely to rise over the following months than at a random starting point. This is a real, documented, tradeable statistical relationship. It is also not remotely a precise timing signal, and the gap between those two facts is where most retail misuse of COT data occurs.

The Specific Misread

Here is the mechanism of the mistake, stated plainly: commercials can be net long — correctly, as it later turns out — for weeks or months before price actually turns. A trader who sees "commercials are extremely long" and treats it as "the bottom is in this week" will frequently be both right about the ultimately-correct direction and wrong about the timing, which in a leveraged short-term position is often enough to be stopped out or drawn down before being proven right.

Suppose commercial net-long positioning in a commodity reaches its highest level in three years in March. Historically, that reading has preceded a meaningful price rise within the following two to eight months — a wide window. A trader who opens a large position expecting the move "any day now" is making a claim the data never supported. The data supported a directional lean over a multi-month horizon, not a precise entry trigger.

Division of Labor With the Rest of the Library

Topic	Where it's covered
Exchange price/quote data (real-time or delayed)	Not covered anywhere in this library — licensed data, never republished
13F institutional holdings	Referenced in passing across several books, no dedicated treatment
COT positioning specifically	This chapter — the first dedicated treatment in the library

No other book in this library gives COT data a dedicated treatment, despite it being free, public, weekly, and — used correctly — genuinely informative. This chapter fills that measured gap directly, rather than continuing to leave it implicit.

Executable Trading Rules

1. Use COT extremes as a multi-month directional lean, never a trigger. A record commercial position is a reason to expect a move is more likely over the coming months, not a reason to enter a trade this week expecting immediate confirmation.

2. Measure "extreme" against the instrument's own history, not an absolute number. A commercial net position that is extreme for wheat may be entirely normal for crude oil. Compare each reading to that specific market's multi-year range, never across markets.

3. Never combine a COT-extreme thesis with tight, short-duration risk management. If the signal itself operates on a multi-month horizon, a stop or position size calibrated for a multi-day thesis will frequently exit the trade before the slower signal has time to play out — the position sizing must match the signal's actual timescale.

4. Treat "commercials are extreme and price is finally moving" as the actual entry trigger, not the extreme reading alone. Waiting for price confirmation on top of the positioning extreme costs some of the move but removes most of the false starts.

Relevance to a Retirement Portfolio

Nothing in this chapter is retirement-account trading advice, and it should not be read as such. COT-based positioning views concern short-term futures speculation, categorically outside a retirement core.

What transfers is narrower and more general: the discipline of matching a signal's actual timescale to the decision built on top of it. A retirement investor evaluating any piece of institutional or macro data — a Fed policy signal, a valuation metric, a sentiment extreme — should ask the same question this chapter asks of COT data: over what horizon has this signal historically been informative, and does my intended action match that horizon, or a much shorter one? Misreading a slow signal as a fast one is the same error whether the account is a futures speculation account or a long-term retirement portfolio being tactically nudged on the strength of a headline.

Chapter 3 covers a related trap: patterns that look statistically real specifically because they were found by searching a small, convenient sample.