How Many Trades Do You Actually Have On?

Here is a book that looks sensible. Long gold. Long euro. Long Aussie. Short dollar-yen. Four instruments, four separate entries, each one sized to the same modest slice of the account, each with its own stop. Four trades, four ideas, risk spread about as evenly as a retail account can spread it.

It is one trade. Every one of those four positions makes money if the dollar weakens and loses money if the dollar strengthens. They were entered separately, they sit on separate lines in the platform, and they will all be right together or wrong together. The account is not carrying four small risks. It is carrying one large one, expressed four times, and the person holding it has no idea because the platform counts tickets and nobody counts opinions.

This is the most common way a careful trader ends up with a position they never decided to take. Not by breaking the risk rule — by keeping it, four times, on the same idea.

Risk-per-trade quietly assumes the trades are different

Almost every risk rule a retail trader is ever given is a per-trade rule. Risk a fixed small fraction of the account on any one idea, so that no single trade can do real damage and a losing run stays survivable. It is good advice and it works, but it carries an assumption in its wording that nobody says out loud: that a trade is a thing that can be wrong on its own.

When four positions share a driver, they cannot be wrong on their own. The move that stops out the first one is the move that stops out the other three, usually inside the same hour, sometimes inside the same minute. What was designed as a per-trade ceiling has become a per-idea floor that is four times higher, and the loss that arrives is not the loss the rule was written to permit.

The rule is not broken. It is being applied to the wrong unit. It was meant to cap what one bet can cost you, and the unit it is being measured against is the ticket.

The one-news test

There is a test for this that takes about thirty seconds and needs no data, no software and no correlation matrix. Look at everything you have open and try to finish this sentence in one go: the single piece of news that would hurt all of these at once is ___.

If you can fill that blank easily — a hot inflation print, a hawkish central bank, a risk-off headline — then that sentence is your actual position, and the tickets are just the way you happened to express it. If you genuinely cannot fill it in, if hurting one position would help another, then you are holding something closer to the several-separate-bets book you thought you had.

The tickets are not the position. The sentence is the position.

Most people, doing this honestly for the first time on a book they are already holding, find they can fill in the blank without pausing. That is the point of the test. It is not hard, and its difficulty is not what makes it useful — what makes it useful is that nobody performs it, because the platform presents four rows and four rows look like diversification.

Correlation belongs to the driver, not to the pair

The usual objection at this point is that the instruments are not that correlated, and there is often a number attached — someone has looked up how two markets have moved together over the last year and found the relationship loose enough to ignore.

That number is an average taken across a period containing many different drivers. During those months, sometimes the dollar was the story, sometimes it was a local central bank, sometimes it was a commodity story that had nothing to do with either. Averaged over all of it, two markets can look substantially independent. That average is a fair description of a typical year and a poor description of any particular hour.

Because correlation is not a fixed property of a pair of instruments. It is a property of whatever is currently moving them. When one driver dominates — an inflation release, a rate decision, a risk event large enough to push everything else off the screen — instruments that normally go their own way move together, and they do it precisely when you have the most at stake. The relationship tightens exactly at the moment the loose historical number told you not to worry.

This is the same mechanism as the release-day pieces on this blog, seen from the portfolio side. A big release does not just make one instrument jumpy. It temporarily makes one driver the only driver, and while that is true, a book that is nominally spread across four markets behaves like a single position with four times the size.

Three drivers that do most of the work

You do not need a taxonomy for this, but it helps to know the handful of sentences that most retail books collapse into.

• The dollar. Almost everything a retail trader touches is quoted against it, so when the dollar is the story — inflation data, a rate decision, jobs figures — most of the screen becomes one bet with different labels on it.
• Risk appetite. Indices, higher-beta currencies and crypto tend to be read as expressions of the same willingness to hold risky things, and on days when that is what is being repriced, they take their instructions from the same headline.
• The instrument's own story. Something specific to that market and genuinely unrelated to the other two. This is the driver that produces real independence, and it is rarer than the number of instruments on your screen suggests.

Gold is worth a note here because it is routinely mis-filed. It is often described as a haven that should offset risk-on positions, and sometimes it behaves exactly like that. On other days it trades as a dollar-denominated asset and falls alongside the very positions it was supposed to balance. Neither behaviour is the true one. Which of them you are getting depends on which driver is live, and that is a question about today rather than a property you can memorise.

Group before you open the last one, not after

The practical version of all this costs a minute and happens before the entry, not during the post-mortem.

• Write one sentence per open position: what has to be true for this to work. Not the setup, not the pattern — the thing about the world.
• Write the same sentence for the trade you are about to take.
• Sort the sentences. Any two that say the same thing are one trade, regardless of what the platform shows.
• Count the distinct sentences. That is how many trades you have on, and if the new one adds no sentence, it adds no diversification — only size.

Doing this before the entry matters more than it sounds. Afterwards it is an explanation; beforehand it is a decision, and it is the only point at which the information can change anything.

A budget per driver, not only per trade

Once you are counting drivers, the risk rule needs a second number. Alongside the ceiling on what one trade may cost you, set a ceiling on what one driver may cost you. Then a book with four positions on a single sentence has to resolve itself one of three ways.

• Split the driver's budget across the four, so each is a quarter of the size it would otherwise be. The book stays wide and the concentration is priced in.
• Take the single cleanest expression of the sentence at full size and skip the other three. Usually the best answer, and the hardest, because three setups you liked go untaken.
• Accept the concentration deliberately and write down that you accepted it, including what it costs if the sentence is wrong.

All three are defensible. The fourth option is the one almost everyone takes: full size on each, no decision made, the concentration discovered on the day it is expensive. That is not a risk appetite. It is an accounting error.

The hedge that is not a hedge

A tempting shortcut is to notice the concentration and neutralise it by opening something in the opposite direction on a correlated instrument. It rarely does what it promises. You are now paying the spread on both, financing both overnight, and holding a position that is neither flat nor directional — its behaviour during the event you were worried about depends on a relationship that, as above, is not stable and is least stable during the event.

If the honest problem is that you have too much exposure to one sentence, the honest fix is to hold less of it. Closing part of a position is unglamorous and it works exactly as advertised, which is more than can be said for an offsetting trade whose offset is an assumption.

What this means if you follow a signal channel

This is where it stops being general advice and starts being specific to how most people actually get their trades. A channel posts several setups in a session. They arrive as separate messages, each complete, each with its own levels, and they read as a menu of independent opportunities.

Often they are not independent, and the reason is structural rather than careless. The setups were found on the same morning by someone looking at the same market conditions, and what made all of them look attractive at once was frequently the one thing that was moving everything. A morning where the dollar is under pressure is a morning that generates several long-gold, long-currency ideas, because that is what the condition produces. The calls are separate. The sentence underneath them is the same sentence.

So taking every call posted in a session is not spreading risk across several ideas. It is increasing size on one idea by way of several tickets, while feeling like the opposite. And a channel cannot correct this for you, for the same structural reason it cannot tell you your position size: it does not know what else you are holding, what your budget is, or which of its calls you already acted on. The grouping has to happen at your end, because your end is the only place the whole book exists.

The practical answer is the second option from the budget list. Read the session's calls, work out how many distinct sentences they contain, take the cleanest expression of each sentence, and skip the duplicates. Then log the ones you skipped and what they did afterwards — otherwise, in a month, you will have no way of telling whether the filter earned anything or quietly cost you a good trade every week.

Common questions

How correlated is too correlated?

There is no number to give you here, and a number would be a false comfort. Any figure you can look up is an average over a window containing many different drivers, which is precisely the information that goes missing at the moment you need it. The question worth answering is not how these two markets behaved on average last year — it is whether the same piece of news would hurt both of them this week. That is the one-news test, and it is about the next event rather than the last hundred.

Then how many positions should I hold?

Also the wrong unit. Count drivers, not tickets. Two positions on two genuinely different sentences is a wider book than six positions on one, and a trader holding those two is carrying less concentration risk than the one holding six, whatever the platform's position count implies. If you want a rule of thumb that is actually about the thing, cap the sentences, not the rows.

Doesn't gold hedge my index positions?

Sometimes, and sometimes it does the opposite, which is why it cannot be relied on as a standing hedge. Gold trades as a haven on some days and as a dollar asset on others, and the label it carries in general commentary does not tell you which one is operating this morning. Ask which driver is live instead. If the day's story is a rate decision, gold and your index position may well be reading from the same script; if it is a geopolitical headline, they may not.

Is crypto separate from all this?

It has genuine drivers of its own, and it also trades as an expression of risk appetite a good deal of the time. Which of those is dominant varies, so it gets the same treatment as everything else rather than an exemption: write the sentence, and see whether it is a sentence you already have on.

Does this still matter if I trade small?

Yes, because concentration is a ratio rather than an amount. Four positions on one driver is four times your intended risk on that idea whether the intended risk was large or small. Trading small changes what the mistake costs; it does not stop it being the mistake, and the habit is the thing that scales with the account.

Why we published this

We publish free calls on Telegram — entries, stops and targets, with the outcome of each one posted afterwards, losing trades included — and we run a paid room and say so openly. Our commercial interest is in you taking the trades we post, and this article argues for taking fewer of them.

It argues for it because the alternative is a reader who takes all four calls from a session, believes they are carrying a diversified book, and finds out on a bad afternoon that they were carrying one position at four times the size they had agreed to. That outcome is worse for them than a skipped trade, and it is not made better by the calls themselves having been good ones. A set of individually sound ideas that all depend on the same thing being true is still one bet.

There is a way to check this on us rather than take our word for it, and it is the reason the free feed is public. Look at the calls posted on a single morning and work out how many distinct sentences they contain. Then look at what happened to them, which is posted either way. If a session's calls tend to win together and lose together, they were one idea, and you should size them as one idea no matter how many messages they arrived in. That is checkable on any channel that publishes its outcomes, ours included, and it tells you more than any description a channel writes about itself.