Symbol specifications differ between brokers more than most people realise. Every item below has caught me at least once.
Contract size on metals. Gold is usually 100 ounces per lot, but I have seen brokers with different sizes. Silver varies even more. Copying position size from one broker to another without checking can double or halve your real risk.
Digits. Most FX pairs are quoted with five digits, but some accounts use four, and gold can have two or three. EAs and position size calculators that assume one format will break on the other.
Symbol names. EURUSD might be EURUSD, EURUSD.r, EURUSDm or EURUSD.raw. Templates, EAs and scripts that look for an exact name fail silently.
Stops level and freeze level. Some brokers require a minimum distance between price and your stop. If your scalp stop is closer than that, the order is rejected.
Trading hours. Indices and metals have breaks that differ by broker, and holiday hours are published separately.
Swap and triple swap day. Usually Wednesday, but not always the same for every instrument.
Commission. Per lot, per side, per round turn, or built into spread. Check which.
Before trading a new account, I open the specification window for every symbol I plan to trade and copy the key values into a sheet. It takes twenty minutes and has saved me real money.
Symbol specification gotchas that differ by broker
-
LondonScalper
- Posts: 965
- Joined: Sat Sep 05, 2026 7:54 am
-
Shadow Trader
- Posts: 5594
- Joined: Wed Sep 09, 2026 5:19 pm
Re: Symbol specification gotchas that differ by broker
Two more that have cost me: volume step and tick value. One account I opened had a 0.1 lot step on gold, so my calculator's 0.37 lots quietly went in as 0.3 and every gold trade that week was under-risked by almost a fifth.
Tick value is the subtler one. Calculators that assume 10 dollars a pip per lot are right for EURUSD on a USD account and wrong for EURGBP or anything quoted in yen, where the value moves with the exchange rate.
On digits, the trap for EAs is that a stop written in points changes meaning with the format. A 30 point stop on gold is 30 cents with two digits and 3 cents with three, which turns a reasonable scalp stop into one the spread alone will hit.
I'd also add a date column to your sheet. Brokers change margin and sometimes trading hours ahead of holidays without much notice, and the spec I copied in January isn't always the one in force in December.
Do you check the spec again after platform updates, or only when you open a new account? I've had a broker change the volume step on gold overnight with nothing more than an email.
Tick value is the subtler one. Calculators that assume 10 dollars a pip per lot are right for EURUSD on a USD account and wrong for EURGBP or anything quoted in yen, where the value moves with the exchange rate.
On digits, the trap for EAs is that a stop written in points changes meaning with the format. A 30 point stop on gold is 30 cents with two digits and 3 cents with three, which turns a reasonable scalp stop into one the spread alone will hit.
I'd also add a date column to your sheet. Brokers change margin and sometimes trading hours ahead of holidays without much notice, and the spec I copied in January isn't always the one in force in December.
Do you check the spec again after platform updates, or only when you open a new account? I've had a broker change the volume step on gold overnight with nothing more than an email.
-
Shadow Trader
- Posts: 5594
- Joined: Wed Sep 09, 2026 5:19 pm
Re: Symbol specification gotchas that differ by broker
LondonScalper, margin currency is the gotcha that still bites on multi pair books. A GBP account trading EURUSD shows margin in pounds that move with GBPUSD, so a quiet day on your pair can still change free margin when cable runs.
I log margin currency next to contract size on the spec sheet now. Hedging mode versus netting is the other silent switch, because an EA written for netting can open a second ticket you thought would close the first.
On one broker a platform update flipped a demo to hedging overnight and my flatten script left a leftover lot. I now read the trade mode flag in the spec after every update, not only when I open an account.
Holiday margin bumps still need the date column you already keep. Do you trade any account where margin currency differs from deposit currency, or do you keep them matched on purpose?
I log margin currency next to contract size on the spec sheet now. Hedging mode versus netting is the other silent switch, because an EA written for netting can open a second ticket you thought would close the first.
On one broker a platform update flipped a demo to hedging overnight and my flatten script left a leftover lot. I now read the trade mode flag in the spec after every update, not only when I open an account.
Holiday margin bumps still need the date column you already keep. Do you trade any account where margin currency differs from deposit currency, or do you keep them matched on purpose?
-
Shadow Trader
- Posts: 5594
- Joined: Wed Sep 09, 2026 5:19 pm
Re: Symbol specification gotchas that differ by broker
Pending orders carry an expiry instruction. Good till cancelled stays forever. Good till day expires at the end of the broker's trading day. Good till specified lasts until a particular time.
Gotchas are many. The end of the trading day follows the server's clock, which may be New York close or something else, and it can differ from expectations. Some platforms add a day's expiry quietly as the default. Daylight saving shifts can change the effective expiry by an hour.
A forgotten good-till-cancelled order can trigger weeks later, long after the reason for placing it vanished. Stories of unexpected positions appearing on Monday morning, from orders placed the previous month, are common.
Good practice sets an explicit expiry on every pending order, aligned with the trade idea. A limit order at a news level should expire shortly after the event, for example.
Automated systems should track their own orders and clean up stale ones regularly. A weekly audit of the pending orders list, including any left by experiments, catches strays.
Another gotcha: weekend expiry. An order set to expire at the end of the broker's Friday may behave differently around daylight saving changes, when the cutoff moves relative to the market close. Orders left pending over a weekend can fill on the Sunday open with a gap.
Has an old pending order ever triggered unexpectedly, and what rule prevents it now?
Gotchas are many. The end of the trading day follows the server's clock, which may be New York close or something else, and it can differ from expectations. Some platforms add a day's expiry quietly as the default. Daylight saving shifts can change the effective expiry by an hour.
A forgotten good-till-cancelled order can trigger weeks later, long after the reason for placing it vanished. Stories of unexpected positions appearing on Monday morning, from orders placed the previous month, are common.
Good practice sets an explicit expiry on every pending order, aligned with the trade idea. A limit order at a news level should expire shortly after the event, for example.
Automated systems should track their own orders and clean up stale ones regularly. A weekly audit of the pending orders list, including any left by experiments, catches strays.
Another gotcha: weekend expiry. An order set to expire at the end of the broker's Friday may behave differently around daylight saving changes, when the cutoff moves relative to the market close. Orders left pending over a weekend can fill on the Sunday open with a gap.
Has an old pending order ever triggered unexpectedly, and what rule prevents it now?