Tradovate CSV Not Importing: Which Report You Actually Need

Aurafy is a young, independent trading journal. It earns trust by keeping the local app usable without an account, making desktop data exportable, and being clear about which features are free and which require Pro.

Page updated September 2026

The four files, by their first line

An account snapshot starts Total P/L, Open P/L, Net Liq, Total Margin Used, Available Margin. It is one row describing your account right now and contains no trades at all. Performance starts symbol, _priceFormat, _priceFormatType, _tickSize, buyFillId, sellFillId, qty, buyPrice, sellPrice, pnl. It holds completed round trips. Orders carries one row per order with a status, a fill price and your stop and limit prices. Fills carries one row per execution. Orders is the best of the four, because the stop price on it lets the journal work out the R-multiple of every trade with no input from you.

The Performance file has no direction column

Nothing in Performance.csv says long or short. Direction is implied by which fill came first: bought before sold is a long, sold before bought is a short. It also uses camel case timestamps that a generic column matcher never matched, so both the direction and the times were lost. The result was one of the worst outcomes in this whole cluster: one user had 105 of his 108 trades stored with every entry and exit time set to the second he uploaded the file, all marked long when most were short. There is now a dedicated parser for this file that reads direction from the fill order and keeps the wall clock times exactly as written. If your Tradovate history looks like that, delete those trades and import again.

The mid-session export gap

The Orders export is a snapshot of what has happened so far. Export it while you are still holding a position and everything after that moment is simply not in the file, including the exit. This produces a correct import that disagrees with your broker, which is maddening to debug because nothing is wrong. One real case: a journal showing minus 290 against a platform showing plus 316, entirely explained by five contracts still open at the moment of export whose winning exit came minutes later. Export flat, at the end of the session, and the gap disappears. If your numbers are off and you exported mid session, check that before suspecting anything else.

Fees, which is the other reason the numbers differ

The Tradovate orders export carries no commission column, so fees have to be estimated. The default is well under what a prop firm simulated account really charges, and the difference compounds fast at any real volume: one account showed 463 dollars of profit against the broker's 106, and the entire 357 dollar gap was commission. The arithmetic that settles it is simple. Take your gross profit, subtract what the broker says you made, divide by the number of contracts you traded, and you have your true per contract rate. Set it on the account, re-cost the existing trades, and the journal matches to the cent.

Paper accounts and multiple funded accounts

Tradovate paper reports export and import identically to live ones. Flag the account as paper in the journal so simulated results do not mix into your real statistics, which is the single most common way a set of trading stats quietly becomes meaningless. The same applies across funded accounts: keep each one separate rather than importing everything into a single default account, because a per account commission rate is the thing that makes each one match its own broker, and a shared account can only carry one rate.