Glossary

How ideas are tested

Backtest

A test of trading rules on past prices. It shows what would have happened, not what will. Every score on this site is a backtest result, not real trading. What a real gold EA backtest should include

Out-of-sample

Prices a strategy was not tuned on. Only results on such prices count here. How ideas are tested

Walk-forward

Tune on one stretch of prices, judge on the next stretch the strategy has never seen, move forward, repeat. Here: tuned on two years, judged on the following six months, eight times, covering 2022 to 2025. How ideas are tested

Overfitting

Tuning a strategy's rules and settings until they fit past prices so closely that they describe that history's noise rather than a lasting pattern. It looks great in the backtest and fades on new prices. Why gold EAs pass the backtest and fail live

Score

The yearly return divided by how much the return swings from year to year, after all costs, counted only on prices the idea had never seen (in finance, the annualised Sharpe ratio). A higher score means a steadier result, not a bigger one. The record

Pass mark

The line an idea has to reach on a test. For the score it is 1.00 after costs, and 0.50 with every cost doubled. An idea has to pass all nine tests, and the pass marks are the same for every idea. How ideas are tested

Luck bar

The score the best of all the ideas tried would be expected to reach by chance alone. Every new idea tested raises it. The record

Never switched on

In no two-year stretch did any setting of the idea make money, so it was never allowed to trade in the test. The record

Enough trades

At least 200 trades in the judged half-years. A result built on a handful of trades says little, however good it looks. How ideas are tested

Costs doubled

The whole test runs a second time with every cost doubled, and the score must still reach 0.50. It catches an edge that only exists when trading is cheap. How ideas are tested

Profitable half-years

At least 5 of the 8 judged half-years must make money, and no single half-year may carry more than half of all the profit. Both catch one good spell carrying the bad ones. How ideas are tested

Neighbouring settings

Settings close to the chosen ones, such as a 19- or 21-day average next to a 20-day one. At least 6 in 10 of the settings tried must also make money, so a result does not rest on one magic number. How ideas are tested

Timing test

The idea has to do better than 99% of 2,000 shuffled copies of itself. It guards against profit that comes from gold rising, not from when the trades were placed. How ideas are tested

Look-ahead (peeking)

A rule that secretly uses a price it could not have known yet. Each idea is run again with future prices cut off, and the results must be identical. How ideas are tested

Trading costs

Spread

The gap between the buy (ask) and sell (bid) price, paid on every trade. It widens at the daily rollover and around news. How ideas are tested

Commission

The broker's fee per trade, charged on many raw-spread accounts. How ideas are tested

Slippage

The difference between the price an order expected and the price it was filled at. On gold it grows around news and in thin hours. How ideas are tested

Swap (financing)

The overnight charge or payment for holding a position past the daily rollover at 17:00 New York time. On MT5 it is often tripled on Wednesday nights. Why gold EAs pass the backtest and fail live

Reading a track record

Verified track record

A trade history a tracking service has read straight from the broker account, usually through an investor (view-only) password, so the trades shown are the trades that happened. How to verify a gold EA's track record

Trading privileges

A separate check that the person showing the record controls the account. History without privileges tells you the trades happened, not that they belong to the seller. How to verify a gold EA's track record

Demo and live account

A demo account trades practice money at the broker; a live account trades yours. Fills, spreads, slippage and pressure all differ, and real results are usually worse. Risk warning

Equity drawdown

The largest fall in account equity from a peak, open trades included. How to verify a gold EA's track record

Partial close

Closing part of a position and leaving the rest open, for example taking profit in three steps. How to verify a gold EA's track record

Profit factor

Money made on winning trades divided by money lost on losing ones, after costs. Above 1.00 the trades made money overall; it says nothing about how steady that was. Research log

Deepest fall

The largest drop from a high point to a later low in a test's results, on the half-years the idea had never seen. Real trading usually falls deeper than a test. Research log

Win rate

The share of trades that made money. On its own it says little: a strategy can win nine trades in ten and still lose money if the tenth loss is big enough. No idea here is judged by it. The playbook

Trading words

XAUUSD

Gold priced in US dollars: the symbol most MetaTrader 5 brokers use for it. Markets

EA (Expert Advisor)

A program that trades on its own inside MetaTrader. How to spot a fake gold EA

Lot

The unit of trade size. On most MT5 accounts one lot of XAUUSD is 100 ounces, so a $1 move in the price is $100 per lot. Check your broker's contract size. Position size calculator

Leverage

Trading a position larger than the money in the account, so a small move in the price makes a large change in the account. Losses can be fast. Risk warning

Stop-loss

An order that closes a losing trade at a set price. Placed with the broker, it limits a loss in normal conditions, and can be jumped over when the price gaps. Position size calculator

Pip

The smallest price step traders quote. Brokers define it differently on gold, as one cent or ten cents, so read the contract specification before you use pips in a formula. What a pip is worth on gold (XAUUSD)

Margin

The deposit a broker holds while a position is open. It is not a fee and not the price of the trade: the gain or loss is on the full position. Gold margin and leverage: what one lot of XAUUSD ties up

Risk reward ratio

How much a trade aims to win compared with what it risks. Aiming for twice the stop distance is a two to one ratio, but the ratio says little without the win rate. Risk reward ratio explained, with the win rate you need

Trailing stop

A stop-loss that follows the price at a set distance while the trade moves your way and stays where it is when the price turns back. Trailing stop on gold: how it works in MT5

CFD (contract for difference)

A contract with a broker that pays the difference between the price at open and at close. You do not own the gold. The costs are the spread and any overnight swap. Gold CFD, ETF, futures or physical gold: how they differ

Martingale and grid

A martingale makes the next trade bigger after a loss; a grid opens more trades as the price moves against the position. Both hide losses until a large one arrives. The rules here allow neither, nor adding to a losing trade. How to spot a martingale gold EA before it costs you

Market data

Volatility

How spread out the daily moves were over the last 20 trading days, scaled to a year. 20% a year is about 1.26% on a normal day. On this site it is realised volatility, from past prices. Volatility

ATR

Average true range: ATR 14 is the average high-to-low range of the last 14 days, in price. The position size calculator can set the stop from it. Volatility

COT report

The Commitments of Traders report: every Friday it shows what big speculators, hedgers and small traders held in each futures market on the Tuesday before. Who holds what (COT)

Large speculators and hedgers

Large speculators (hedge funds, trend followers) trade for profit, and their net position is the report's headline number. Hedgers (miners, producers, banks) usually take the other side. Who holds what (COT)

Crowded

A market near 0 or 100 on its 3-year range of positions: when that many have already bought or sold, a turn can be sharp. It is a reading of positions, not a timing signal. Who holds what (COT)

Fed rate odds

How likely a cut, a hold or a hike is at the next Federal Reserve meeting. Here it is estimated from US Treasury bills, so an outcome under about 20% means no strong view. Fed rate odds

Yield curve

The interest rates on US government debt from the shortest bills to 30-year bonds, drawn as one line. When short rates sit above long ones the curve is called inverted. Fed rate odds

Jobs report

The monthly US count of jobs added and the unemployment rate, usually on the first Friday of the month at 8:30 New York time. One of the releases that move gold most. Economic calendar

CPI

The consumer price index: the monthly US measure of how fast prices are rising (inflation), out at 8:30 New York time. Economic calendar

PCE

The personal consumption expenditures price index: a monthly US inflation measure, and the one the Federal Reserve aims at. Economic calendar

GDP

Gross domestic product: the value of everything the US economy produced, published each quarter, first as an estimate and then revised. Economic calendar

Crypto futures

Perpetual future

A futures contract with no end date, traded mostly on crypto. Funding payments between buyers and sellers keep its price near the coin's price. Crypto futures

Funding rate

What one side of a perpetual future pays the other. Positive: buyers pay sellers, so more traders are long with borrowed money. A high rate for days means a crowded long; a negative rate, a crowded short. Crypto futures

Open interest

The value of all futures positions still open. Rising with the price: new money is behind the move. Falling sharply: positions are being closed or forced out. Crypto futures

Puts against calls

Open put options divided by open call options. Above 1, more open options pay off if the price falls (protection is being held); below 1, more pay off if it rises. Crypto futures

DVOL

Deribit's index of how much options traders expect a coin to move over the next 30 days, as a yearly %. 50 is about 2.6% up or down on a normal day. Crypto futures

Fear & Greed Index

One number for the mood in crypto, read from volatility, volume, social media and Bitcoin's share of the market. Extremes are where crowds have already acted, not a timing signal. Crypto futures

Something missing or unclear? Tell us.