A commodity trade is agreed in a few minutes, and the price is what everyone is watching. It is also the term that causes the least trouble afterwards - both sides wrote it down and meant the same thing by it. The tolerance, the governing measurement, the payment terms and whatever "usual terms" resolved to were settled in the same minute, with a phrase.
Summary
A commodity trade is agreed in a few minutes, and the price is what everyone in those minutes is watching. It is also, reliably, the term that causes the least trouble afterwards. Both sides wrote the price down, both sides meant the same thing by it, and it is a single unambiguous number. Everything else agreed in that same minute — the tolerance and whose option it is, the quality specification, which measurement governs, the payment terms, what "usual terms" refers to — is settled faster, written more loosely, and is where the money is later found to have gone.
This article sets out what actually happens at the moment of execution: how wide the negotiable surface really is, why agreeing terms by reference imports a document nobody re-read, why the person who writes the recap is choosing the words the deal will be remembered by, and why a term the trading system has no field for is a term that will not reach settlement. It is written for the people on both sides of that minute — traders who execute and then live with what they agreed, the middle office who has to turn a conversation into a record, and the operations and settlement desks who find out months later which terms were captured and which were assumed.
If you want the mechanism, start with The price gets the attention, and the rest of the deal gets the same minute. If you already execute daily and want the part that fails quietly, start with A term with no field will not reach settlement.
The price gets the attention, and the rest of the deal gets the same minute
Price is one term among many, and it is the only one that reliably receives full attention at the moment it is agreed. The rest are agreed in the same breath, by people whose minds are on the price.
Think about what actually has to be settled before a physical trade exists. Quantity, and the tolerance around it, and which party may exercise that tolerance. The delivery period, and who may narrow it. The quality specification, and what happens if the cargo is off-spec rather than out of spec. Which measurement governs, and whose inspector performs it. Payment terms, and whatever security stands behind them. Governing law. The demurrage rate. Each of these carries money, and several of them carry more money than a modest move in the price.
They are not negotiated with equal care, and it would be strange if they were. The price is the number both parties are being judged on, it moves while you talk, and the whole conversation is organised around it. The other terms get resolved with a phrase — usual terms, same as last time, standard for that grade — because resolving them properly would take longer than the price will hold.
That is not carelessness. It is the correct response to a market that does not wait. But it produces a pattern that follows directly from it: the term everybody concentrated on is the one whose meaning is not afterwards in doubt, and the terms settled in a phrase are the ones whose meaning has to be established later, at a point when both sides have an interest in the answer. Disputes gather where meaning is uncertain. That is the argument of this article, and whether it describes your own last several disputes is something your files can settle better than we can.
"Usual terms" is a pointer, and pointers rot
Agreeing by reference does not copy a set of terms into your trade. It points at a document, and the document it points at was written for a different deal.
Our usual terms means whichever terms the two parties last used, or a house set of general terms and conditions, or a previous contract that someone will go and find. Standard for that grade means whatever the market convention is understood to be, by two people who may understand it differently. In each case the substance of the agreement lives somewhere else, and the trade carries only a reference to it.
References have three properties worth stating plainly, because none of them is visible at the moment the reference is made.
They resolve differently for each party. Your usual terms and theirs are two different documents. If the last deal between you was papered on their form, "usual" may now mean their form — a fact that only becomes apparent when somebody goes to look.
They import everything, including the parts that were negotiated hard the last time and the parts that were conceded. A clause that one side accepted reluctantly in a previous trade, under different market conditions, arrives in this one as a settled matter.
And they go out of date without anyone editing them. House terms get revised. A grade's market convention shifts. The counterparty's standard form changes when their legal team updates it. The reference stays the same while what it points at moves, which means a trade agreed on identical words in January and in June can be two different trades.
Whoever writes the recap chooses the words
Between the moment a trade is agreed and the moment a contract exists, the operative text is a short message written in a hurry — and whoever writes it has chosen the wording that everything downstream will be built on.
The recap is the message a trader or a broker sends immediately after agreement, setting out what was just agreed. In practice it is the trade until a contract is signed and often for a good while after, because it is what the middle office types from, what operations plans against, and what both parties will read again if there is ever a question.
It is written by one side. That side is not being sharp; they are being fast, and they will naturally write the version of the conversation they remember. But two things follow that are worth being awake to.
Ambiguity written by you reads as clear to you. The author is the person least able to notice that a phrase has two readings, because they know which one they meant. The counterparty reading it has a much better chance of spotting the ambiguity — and, if the ambiguity happens to favour them, the least incentive to raise it.
Silence in a recap is a term. A recap that does not mention who holds the quantity tolerance has not left that question open; it has left it to be resolved later, by whoever has the stronger document or the stronger memory. The absence of a term is not neutral, and it is invisible in exactly the way that makes it hard to review.
The practical form of this is not fraud and rarely even friction. It is that a term everyone believes was agreed turns out never to have been written by anyone, and the first person to discover this is the one who needed it.
An option granted in a sentence is an option nobody charged for
Several of the terms agreed at execution hand one party a choice. A choice has value. The price agreed in the same minute is not negotiated as though it does.
The tolerance around a quantity is an option: whoever may exercise it will exercise it when it suits them, which is when it does not suit the other side. A delivery period that one party may narrow is an option on timing. A range of discharge ports from which one party may select is an option on freight and on destination economics. A specification band within which a cargo may be delivered is an option on quality.
None of these are exotic. They are ordinary commercial terms, they exist for good operational reasons, and a trade without any of them would be unworkably rigid. The point is narrower and it is about pricing rather than about drafting: the party who grants the option is short something, and the price they agreed was almost certainly negotiated as though they were not.
What makes this durable rather than a one-off is that the granting happens in a register where value is not discussed. Price is discussed in money. Terms are discussed in words. The trade is agreed as a single package, and nothing in the process asks which of the words were worth money. So the same option gets given away repeatedly, by different people, without anyone ever seeing an aggregate.
The useful test is not "should we have charged for this", which is unanswerable at speed. It is whether anyone could say, across the book, which options this desk is short and to whom — and if the answer requires reading contracts, then the answer is no.
A term with no field will not reach settlement
A trade record is not a copy of the agreement. It is a translation of the agreement into the fields the system offers, and a term with no field has nowhere to go.
Every trading system has a schema: a set of things it expects a trade to have. Price, quantity, counterparty, delivery period, instrument — these have fields, and they survive. Terms outside the schema have three possible destinations, and only one of them works.
A free-text note. The term survives as prose. It can be read by a person who thinks to look and cannot be reported on, checked, aggregated or alerted against. It is present and inert.
An attachment. The contract or the recap is filed against the trade. This is genuinely better than nothing and is still not a field: nothing downstream can act on a sentence inside a PDF, and the number of people who will open it before it matters is small.
Nowhere. The term was agreed, it was in the recap, and the person entering the trade found no place to put it, so it stopped there. This is the default, not the exception, because entry happens under time pressure and a field that does not exist cannot be filled.
The consequence arrives late and looks like something else. At settlement, a difference appears that traces back to a term nobody disputes was agreed — and the question becomes why the invoice did not reflect it. The answer is not that anyone forgot. The term never became data, so nothing that ran on data could have used it.
This is also why "we should be more careful at trade entry" is the wrong remedy and never works. Care cannot enter a value into a field that is not there. The decision that matters was made earlier, by whoever decided what the trade record consists of — and that decision is a statement about what the business intends to manage, whether or not it was made deliberately.
Under time pressure you cannot check more, only check faster
Execution quality is not decided at execution. It is decided beforehand, by what somebody made possible to look up in seconds.
At the moment of a deal there is time for one or two questions, not for five. That constraint is real and it will not yield to discipline, training or a checklist — a trader who pauses to open a contract loses the trade, and losing trades is also a cost. Anyone whose proposed improvement begins with take a moment to verify has not accepted the constraint.
What does yield is preparation. The questions asked at execution are largely the same questions every time: what did we agree with this name last time, what is our standing position on this term, how much do we already have with them, which of our forms did we use. Each of those has an answer that existed before the call started. The difference between a desk where those answers are one glance away and a desk where they require a search is not a difference in how careful the traders are.
Two consequences follow, and they point in a more useful direction than exhortation.
The scarce resource is retrieval, not judgement. A trader can evaluate a term in seconds if they can see it. The failure is almost always that they could not see it in time, and so used a phrase that deferred the question.
And the terms most worth pre-deciding are the ones that recur. A desk that has decided in advance where it stands on tolerance, on governing measurement, on which form it papers on, has removed those from the minute in which they cannot be considered properly. That is not a system project. It is a decision, written down once, made available.
Where this leaves the middle and the back office
The pattern is the same one that shows up everywhere downstream of a trading desk, and execution is where it starts. The facts that will decide the money are created in a conversation, recorded in a message, and translated into a record by someone who was not in the conversation. Each of those steps is lossy, and the loss is not random: it falls on whatever the system has no field for, which is the same set of terms every time.
The trading desk's own work is well supported. Price, quantity and counterparty go into a system built to hold them and are correct. The terms that decide whether the invoice will be right — the tolerance and whose it is, the governing measurement, what "usual terms" resolved to on this particular occasion — sit in the recap, in an attachment, or in the memory of the person who executed. They are not missing. They are simply not in a form that anything can act on.
The cost is paid twice, in different currencies. Once at settlement, where a difference has to be researched back to an agreement rather than looked up. And once continuously, in the fact that nobody can answer portfolio questions about terms — which options are we short, how many trades carry this counterparty's form, where did we agree a governing measurement we would not choose today. Those questions are not hard. They are unanswerable, which is a different problem and a fixable one.
What is agreed in the same minute as the price
| Term | What it actually decides | Where it usually lives right after execution |
|---|---|---|
| Quantity tolerance, and whose option it is | Who benefits when the market has moved by loading | A phrase in the recap, if it is mentioned |
| Delivery or loading period, and who may narrow it | Which market the cargo is delivered into | The recap, then the contract |
| Quality specification and its tolerances | Whether an off-spec cargo is rejected, discounted or accepted | The contract, by reference to a standard |
| Which measurement governs, and whose inspector | The number that settles, when two numbers exist | The contract, often by reference |
| Payment terms and any security | When cash arrives, and what happens if it does not | Standing terms, assumed rather than restated |
| Demurrage rate and laytime terms | The size of a cost created by events nobody controls | The contract, by reference to a standard |
| Governing law and dispute forum | Where an argument is held, and therefore how it ends | House terms, rarely discussed at execution |
| What "usual terms" resolves to on this occasion | All of the above, when they were not stated | Nowhere — it is a reference, not a record |
The bottom row is the one to sit with. It is the term that determines the meaning of every other row above it, and it is the only one that is not written down anywhere as a value.
What has to leave the desk with the trade
These are fields on a trade record rather than clauses in a contract. Each is known within minutes of execution and gets harder to establish every day afterwards.
| Field | Why it is load-bearing |
|---|---|
| Which document set this trade is on, resolved to a specific version | "Usual terms" is a pointer; this is what it pointed at, on this day |
| The tolerance, and which party may exercise it, as two separate values | One without the other is not usable by anything downstream |
| Which measurement governs quantity, and which governs quality | This decides settlement, and it is cheap now and expensive later |
| Every option granted, recorded as an option rather than as prose | An option in a sentence cannot be aggregated; the same option in a field can |
| The recap itself, linked to the trade and identifiable as the operative text | It is what both parties will read if there is a question, so it should be findable without a search |
| Who executed, and who entered — when they are different people | A later question about what was intended comes back to one of these two |
| Anything agreed that has no field, flagged as such rather than left in a note | A known gap can be worked; a gap inside free text cannot be found |
A trade carrying these seven can be settled by people who were not there. A trade missing them settles correctly whenever nothing unusual happens, which is the ordinary case — and that is exactly why the gap is not noticed until it is expensive.
Questions people ask about this
Why do our disputes never seem to be about price?
Because price is the term that received full attention from both sides, was written down by both sides, and is a single number with one meaning. Almost everything else in the trade was agreed in the same minute with a fraction of that attention, often by reference to a document rather than in words of its own. Disputes gather where meaning is uncertain, and price is the one place it is not.
What does "usual terms" actually commit us to?
Whatever document it resolves to — and the honest answer is that it resolves to different documents for the two parties until somebody checks. It may mean your standard terms, theirs, the terms used in the last trade between you, or a market standard form. It is worth establishing in advance which one your desk means by the phrase, and worth recording on each trade which one it resolved to, because that is the fact that will be needed and the one nobody writes down.
Is the recap a binding document?
Whether it binds is a legal question about your jurisdiction and your particular wording, and it is not the useful question here. The useful question is that the recap is operative in practice: it is what the middle office types from, what operations plans against, and what both parties reread when something is unclear. Treat it as the text that matters, because everybody downstream already does.
Should we be charging for tolerances and options?
Pricing each one individually at the moment of execution is not realistic, and attempting it would cost more deals than it saves. The achievable version is to know what you have given away in aggregate: which options this desk is short, to which counterparties, and in what size. That is a reporting question rather than a negotiation question, and it becomes answerable as soon as options are recorded as options rather than as sentences inside a contract.
Our traders are experienced. Isn't this a training problem?
Experience helps with judgement and does not help with retrieval. The recurring failure at execution is not that somebody evaluated a term badly; it is that they could not see the relevant fact within the seconds available and used a phrase that deferred it. That is a preparation problem — what has been decided in advance, and what can be looked up instantly — and it does not improve with seniority.
Why does trade entry keep losing terms we definitely agreed?
Because entry is a translation into whatever fields exist, and terms without a field have nowhere to go. Somebody entering a trade under time pressure cannot invent a field, so the term ends up in a free-text note, in an attachment, or nowhere. None of those can be reported on. The remedy is a decision about what a trade record should consist of, not more care from the person typing.
We attach the contract to every trade. Isn't that enough?
It is much better than nothing and it is not the same as capturing the terms. An attachment makes a term retrievable by a person who already suspects it exists and knows to look. It does not make the term available to anything that runs automatically — a check, a report, an exposure calculation, an invoice comparison. The distinction that matters is not whether the information is stored but whether it is in a form something can act on.
What is the smallest useful change here?
Pick the two or three terms that have actually caused your disputes — commonly the governing measurement, the tolerance and whose it is, and which document set applies — and give each of them a real field on the trade record. Then require them at entry. It is a small change, it will be unpopular for a while, and it converts a whole category of settlement research into a lookup.
Where this lands in a trading system
Nothing above is an argument for buying software. It is an argument about what a trade record should consist of, and that is a decision rather than a purchase. Some of it does map onto system capabilities, and it is worth being precise about which parts.
The first conclusion is that the terms agreed alongside the price are part of the trade and have to live with it rather than beside it. That is the territory of a CTRM/ETRM system providing complete trading lifecycle management for commodity traders, integrating physical trade, financial hedging, risk control, and settlement in one platform; Time Dynamics' Fusion is one, and the part of it that bears on this article is complete physical trade lifecycle management and execution — the trade and its downstream consequences held as one object, which is the condition under which a term captured at execution can still be present when an invoice is checked.
The second conclusion is the awkward one: if a term has no field, no system will capture it, and adding a field is a decision about what the business intends to manage. Where the terms that matter to a particular desk are not the ones a standard schema anticipated, that is what tailored solutions designed specifically for your unique business requirements addresses. It is worth being clear-eyed about the order of operations, though — the decision about which terms deserve a field comes first and comes from the desk, and no supplier can make it for you.
The third conclusion is that execution quality depends on retrieval: what can be looked up in the seconds available. The facts a trader needs at that moment — what was agreed with this name last time, which form was used, what the standing position on a term is — usually sit in previous contracts, in a counterparty file and in somebody's spreadsheet. Getting them into one place is what X-Ray is for: a non-invasive data processing and analysis platform designed specifically for enterprise clients, whose collection toolkit XDK performs non-invasive automated data collection from databases, Excel files, and web interfaces, and which is built to collect data without disrupting existing systems or workflows — relevant here because the counterparty file is working and nobody is going to stop using it during a trading day.
No system negotiates. It will not tell you a term is unfavourable, will not price an option you are about to grant, and has no view on whether your usual terms are good ones. It does not write the recap, and it cannot recover a term that was agreed out loud and written down by nobody — that term is gone, and the only remedy is upstream of anything on this page. What a system can settle is narrower and duller than the problem in this article: whether the terms someone decided were worth capturing are actually captured, whether they can be reported on rather than read, and whether the person checking an invoice next quarter can see what was agreed without finding the person who agreed it. That is a real part of the gap. It is not the whole of it, and knowing which part you are buying is the difference between a system that helps and one that disappoints slowly.
A note on the evidence in this article
There are no statistics here, and that is deliberate. The observations behind it — which terms get agreed by reference, where a recap sits in practice, what happens to a term with no field — come from material we are not able to publish as a citable figure. A number nobody can check would make this read as more authoritative and would make it worth less.
So the argument runs on mechanism, and it is written to be checked against your own desk rather than believed. Three tests are proposed above and they are the intended use: pull your last several disputes and see how many turned on price, ask what "usual terms" resolved to on a trade agreed last month, and take one term you know you agree regularly and find out whether it exists as a field anywhere. If your own book answers those differently from what is written here, your book is the better evidence.