Note /
Stock draw vs. stock build: why the barrel count isn't the whole story

A stock draw means crude or product inventories fell week over week. A stock build means they rose. That's the whole definition, and if trading were that simple every desk would just read the headline number off the EIA release and call it a day.
It isn't that simple, because the number by itself tells you almost nothing about what caused it.
What a draw or build actually measures
The weekly change EIA reports is a net figure: production plus imports minus exports minus refinery runs, rolled up against what came out of or went into storage. A 4-million-barrel draw could mean refiners ran hard and ate through stocks. It could also mean imports got stuck behind a hurricane and didn't show up that week, with the barrels arriving the following Wednesday as a build that erases the draw entirely. Same storage behavior, two very different weeks on paper.
This is why a draw that beats the API number from the night before, or beats the Street consensus, moves price more than a draw that simply happened. The market prices the surprise relative to what positioning already assumed, not the raw barrel count sitting in the tank.
Seasonality complicates this further. A 2-million-barrel build in late February is bearish on its face but can be perfectly normal for that calendar week, while the same build in July, deep into driving season, is a different signal entirely. Anyone who's traded through a few turnaround seasons knows the five-year average range matters as much as the print itself, and that range shifts every week as old data rolls off.
Why the headline draw can hide the real story
A few things routinely make the size of a draw or build misleading on its own:
Revisions. EIA restates prior weeks almost every release. A draw reported last Wednesday can get revised smaller two weeks later once better balancing data comes in, which means trading the headline without checking whether last week got revised is trading half the picture.
PADD distribution. A national draw can mask a build in PADD 3 (the Gulf Coast, where most of the export and refining capacity sits) offset by a draw somewhere less relevant to the barrels that actually move markets. The national number nets these out, so a flat week nationally can still mean real tightness or real length in the region that matters for your position.
SPR activity and exchanges. Releases from or additions to the Strategic Petroleum Reserve show up in the data and can swing commercial crude stocks without any change in underlying supply-demand balance.
Floating storage and barrels in transit. Crude sitting on a tanker between loading and discharge isn't in anyone's onshore tank count that week. It shows up as a draw at origin and a build at destination days or weeks apart, and the lag between those two entries is exactly the kind of noise that makes a single week's print a weak signal on its own.
Reading the draw or build in context
The weekly number still matters. The useful read comes from comparing it against your own expectation, not just the consensus guess, and from knowing which part of the barrel count you're watching: crude versus gasoline versus distillate versus Cushing specifically.
The structural problem is timing. EIA and API both report once a week, and by Wednesday afternoon the market has already repositioned around whatever the number says, whether or not that number holds up on revision. A desk that's watching tank levels at the terminals that matter, not waiting for the weekly print to confirm what's been happening in storage all along, gets to form a view before the release rather than react to it. That's the gap Oil Inventory Index is built to close, with a daily read on floating-roof tank levels at the facilities you already track.
If you're tired of positioning around a number you can't see coming, a daily independent stocks read is worth a look before the next release.