The grade is a rumour

A trader spent an afternoon teaching us how a mine actually gets built. Then they handed us a project that quietly undid everything I thought a good deposit was.

Before this month I would have told you a good mining project is one with a lot of high-grade metal in the ground. I'd have said it with confidence, because it sounds obviously true, and because nobody had yet shown me the number I was actually looking at.

Two ways into the ground

A trader walked us through how a mine forms, and it starts before anyone has dug anything, back when it's still just land with potential and a choice between the two ways to open the ground.

Open-pit: you strip the overburden and dig downward in benches. It's cheap per tonne and high-volume, and it's what you use when the metal sits shallow and spread thin (low-grade, but a lot of it).

Underground: shafts, tunnels, ventilation, ground support. Far more expensive per tonne, which only pays when the metal is deep and rich enough to chase selectively. The choice isn't taste. It's arithmetic. Depth, grade, and the ratio of waste to ore decide it long before anyone is allowed to have an opinion.

Then the rock has to become a product. Crush it, grind it to powder, float off the valuable minerals, and what leaves the site is concentrate, the stuff that eventually lands near the desk I sit at. I'd been reconciling concentrate for weeks without a clear picture of the machine that makes it. It's a slightly embarrassing thing to admit and a useful one to fix.

Where we actually sit

The part that reorganized my mental map was where Glencore stands in all of this.

Most people picture either a miner or a trader. We're both: an integrated merchant-miner. We own and run mines and smelters, and we run one of the largest physical trading desks in the world. A pure miner lives and dies by the price of the one thing it digs up. A pure trader owns nothing and moves everything. Sitting across both changes what the metal even is to me: not an asset sitting still, but a link in a flow that runs from a hole in the ground in one country to a smelter's queue in another.

That's the whole point of the model. You see the same tonne at every stage, so you can price the friction between the stages, which, if you've read anything else I've written here, is the only lesson I seem able to learn twice.

The number that fooled me

Then came the project. We each had to choose which junior mining projects to cover, and at what stage they sat, then dig through the numbers and write the whole thing up as a summary for the traders. Not would you fund this. Something quieter, and it turned out harder: what here is worth a trader's attention, and how do you say it so they can use it?

My instinct went straight to grade. The assay results. Percent copper, grams of silver. I ranked my candidates by grade and felt quite pleased with myself.

Grade, it turns out, is the least interesting honest number on the page. Because a grade is only as real as the confidence behind it, and the industry is unusually explicit about that confidence. A deposit's metal gets sorted into three categories (inferred, indicated, measured) that are really just a spectrum of how much anyone actually knows.

An inferred grade is a reasonable guess from a handful of drill holes: geologically plausible, statistically thin, and (this is the part that stopped me) not even allowed into the economics of a serious feasibility study. An indicated or measured grade is one that's been drilled tightly enough to bank on.

So I'd been asking the wrong question. Not is the grade high? but is this grade predicted, or is it known? Predictable versus actual. A dazzling inferred number and a modest measured one are not the same asset in different outfits. They're different bets entirely. I'd spent a day ranking rumours by how loud they were.

What a trader actually needs

Once grade stopped being the headline, the exercise rearranged itself around a different question: which of these is even worth writing up, and how far along does it have to be before a trader should care?

Because a mine isn't a single thing. It's a sequence of gates, each one converting a little uncertainty into cost. A scoping study, or PEA, is the first rough economic sketch, allowed to lean on inferred resources and honest about being a sketch. A pre-feasibility study, PFS, is where the numbers get real: engineering choices made, costs pinned to something like a quarter either way. A definitive feasibility study, DFS, is the bankable version, tight enough that a lender will actually lend against it.

The project I led with sat at the PFS stage, and that's what made it worth a trader's time. A deposit still years from a scoping study is a geology story: interesting, not yet actionable. A PFS-stage project has crossed enough gates that its metal, its timeline, and its rough economics are real enough to plan around. That's the difference between something the desk files under someday and something it files under watch this.

And then the actual deliverable: the summary. Writing for a trader is an exercise in subtraction. They don't want the geology essay. They want four things: is it real, when could it produce, what's the metal, and what would move it. And they want the parts I chose to leave out to be the right parts. Half the work of the research was deciding what a busy person didn't need to read. Which, I'm realizing, is most of what a good report is.

A few weeks, four things

• Grade is a claim. The resource classification is whether anyone should believe it.

• "Inferred" is a polite word for we think so. It can't go in a feasibility model, and it doesn't belong at the top of a note.

• The stage decides whether a project is a story or a signal. Surface the ones that have crossed enough gates to be actionable.

• A summary for a trader is mostly subtraction. Deciding what to cut is the research.

I chose what to write up in an afternoon, and I'm fairly sure a real analyst would find the hole in it by lunch. But the thing I think I got right was learning to distrust the biggest number on the page. The grade was always the easy part to admire. What the deposit is actually worth is a slower, duller question about how much of it anyone has bothered to prove. And, as usual around here, the dull question is the one that pays.