My numbers look wrong

You are comparing Alltra against a broker statement, or against what you remember, and they do not agree. Before treating that as an error it is worth knowing that a difference in TRADE COUNT is expected, while a difference in NET P&L on the same trade is not.

How a trade is built

Alltra walks your fills for one instrument on one account in time order, carrying a running position. Going from flat to a position opens a trade. A fill in the same direction adds to it and the average entry is recalculated. A fill in the opposite direction takes some of it off and realises P&L on the quantity it matched. Returning to flat closes the trade.

So one order filled in five pieces is five fills and one trade. Scaling in three times and out twice is one trade with five legs, not five trades. And a single fill large enough to close your position and open the opposite one is split at exactly zero: the first trade closes, a new one opens with what is left. A broker export that prints one row per fill, or that merges a day into a summary, will not agree with that — and neither of you is wrong about the money.

Where the money can genuinely differ

• Fees and commission — a trade’s outcome is derived from its NET P&L, so a gross win that costs more in fees than it made is a loss in Alltra. Commission and fees are recorded separately and shown summed as one cost.
• Computed fees — where a broker reports fees, those are used and never altered. Where it reports none — which is normal for a CSV or a hand-typed trade — Alltra computes them from the instrument’s fee schedule. Nothing on screen yet distinguishes a reported fee from a computed one, so on an imported or hand-typed trade a small difference in the cost is expected rather than an error.
• The contract behind the symbol — futures P&L is ticks multiplied by tick value multiplied by contracts, not a price difference. A micro and a mini contract on the same symbol have different values, so the wrong instrument is wrong by a whole multiplier rather than by a rounding.
• Which entry an exit closed — first-in-first-out, last-in-first-out and weighted average all produce the same TOTAL for a trade and different amounts against each partial exit. The convention is held per account, there is no control for it in the product yet, and every account matches first-in-first-out today.

Where two Alltra surfaces answer differently

A few figures are computed with different denominators in different places, and when two pages disagree over the same trades that is usually why. Win rate is the clearest: the Performance dashboard divides wins by wins plus losses plus breakevens, so a trader who scratches often is not flattered by it, while the daily Report divides wins by wins plus losses and leaves breakevens out. Over a window with no scratches the two agree exactly; over one with several they cannot.

“Win/Loss Ratio” is one label over two different figures. On the Performance dashboard it is a COUNT — winning trades divided by losing trades, read as “X : 1”. The ring of that name on a daily Report is a SIZE — the average winner divided by the average loser. They routinely point in opposite directions: eight small winners against two large losers is a healthy count ratio and a poor size ratio, and that gap is the thing worth reading rather than a mistake.

What to check, in order

1. 1Scope, before anything else — the account picker, the date range, and whether Demo Mode is on. A figure computed over a different set of accounts is not a wrong figure.
2. 2Which day the trade counts on — a trade is filed under the day it CLOSED, resolved in the timezone saved in Settings › Preferences. An overnight position belongs to the day it was exited, and a late-evening trade sits on the wrong day if that zone is wrong.
3. 3Open the trade and read its Executions tab — it lists the real fills the trade was built from, labelled entry, scale in, partial exit and exit. If those fills match your broker, the trade is right and the disagreement is about grouping.
4. 4Compare the cost — the trade’s panel shows Fees — commission and exchange fees summed — and the Trades table has the same figure as a column you switch on under Columns. If the gross agrees and the net does not, that is the whole of the difference, and on a CSV or hand-typed trade the fees were computed rather than reported.
5. 5Check the instrument if it is out by a factor — a P&L that is exactly five, ten or fifty times what you expected is a contract multiplier, not an arithmetic error.
6. 6Check the per-exit split if the total agrees — the total being right while each partial exit is not is the lot-matching convention. Alltra matches first-in-first-out; a broker reporting last-in-first-out or an average splits the same total differently.
7. 7Read what the figure is actually called — profit factor, expectancy and the two win/loss ratios all have precise definitions in the glossary. Several disagreements are two names for two different things.
8. 8Give it a moment after an import or a sync — trades are derived after fills land, and dashboards fold them in after that. A dashboard can honestly be a few seconds behind an import.

Why a wrong number is usually temporary

Because trades are derived rather than typed, they can be thrown away and rebuilt from the same fills. A late correction from a broker, a fee update or a fix to the way a contract is priced marks the affected window and replays it. That is also why a fill Alltra cannot map to a known instrument is never imported quietly — it waits to be mapped rather than being guessed at, because one wrong multiplier corrupts every figure on the account.

When to tell us

When the fills in the Executions tab match your broker and the money still does not. Send one example — the account, the symbol, the day, what Alltra shows and what your statement shows. One concrete trade is worth more than a description of the pattern.