What an order block is
An order block is the last candle against the move before price leaves an area with force: the last down candle before a strong rally, or the last up candle before a sharp drop. The idea behind it is simple. Somebody large enough to move gold was filling orders there, and when price comes back to that candle, some of those orders may still be waiting.
On a gold chart that is easy to draw and hard to use. Gold moves enough in a session to leave dozens of these candles behind, and an order-block indicator that draws every one of them gives you a chart covered in boxes, most of which never do anything. The question is not “where are the order blocks”. It is “which of them are worth waiting for”.
Why most order-block indicators draw too many
A typical script finds every block on one timeframe, keeps it until price touches it, and stops there. Three problems follow:
- Every block looks the same. A block that formed in a quiet Asian session and one that formed on a London breakout get the same box.
- Nothing knows the context. A buy block in a falling market is drawn exactly like one in a rising market.
- Failed blocks stay on the chart. A block that price ran straight through still looks like support until you delete it by hand.
None of this is a coding problem. It is a filtering problem, and it is solved by asking more questions of each block before it is allowed to matter.
The filters that cut them down
These are the filters we found worth keeping on gold. Each one throws blocks away, and that is the point.
1. Two timeframes, not one
Build the library from two timeframes at once, for example the 5-minute and the 15-minute chart, and colour them differently. A 15-minute block seen on a 5-minute chart is a bigger, slower level that the smaller blocks sit inside. Use the higher one for context and the lower one for timing.
2. Mitigated, broken, and old blocks
A block is used up once price has traded back through it. Remove it, or better, turn it into a breaker block: a failed buy block that price broke down through often becomes resistance from the other side. Blocks that have sat untouched for a long time should fade, so the eye goes to the fresh ones.
3. Round numbers
Gold traders watch round prices: the ×100, ×50, ×25, ×10 and ×5 levels. An order block that sits on one of those prices has two reasons to hold instead of one. A block in empty space between round numbers has only its own. Keeping the blocks that line up with a round number and fading the rest removes a large share of the noise in one step. There is more on this in Round numbers and support and resistance on gold.
4. Blocks that formed with intent
Some blocks are left behind by a real move; others are the residue of drift. Score each new block on what happened as it formed: how strong the impulse away from it was, whether it swept liquidity above a recent high or below a recent low first, whether it left a fair value gap, and whether it sits on a round number.
5. Regime: is the market trending at all
A buy block in a clean uptrend and a buy block in a choppy range are different trades. Gate the signals on the market’s state: a higher-timeframe trend direction and a trend-strength reading such as ADX. In a range, most continuation setups should simply not fire.
6. Session
Gold behaves differently across the day. A counter-trend setup that is reasonable in one session can be a poor idea in another. A session filter that blocks counter-trend entries during the hours you don’t want them is cheap and effective. Use a city time zone (Europe/Athens, London, New York) rather than a fixed UTC offset, or the window slides by an hour when daylight saving changes.
7. Don’t buy into the ceiling
A long entry just below an untested daily high, or a short just above an untested daily low, is heading into a level the whole market can see. Refusing those entries removes signals, and it should.
What a signal is, and what it is not
When every filter passes, price has come back to a block where a measured configuration found favourable behaviour in the past. That is a hypothesis, not an instruction. It needs a defined trade before it is anything at all: an entry, a stop past the block, targets in multiples of the risk, and a position size that risks a fixed share of the account.
How to know if your filters work
Filters are easy to invent and easy to fool yourself with. Three rules keep the test honest:
- One position at a time, with cooldowns. An indicator can print overlapping signals; your account cannot take them all. Test what you could actually have traded.
- Measure one configuration at a time. If five rule sets run together, you cannot tell which one produced the result.
- Keep a window you never tuned on. A configuration that only works on the data that chose it has not been tested. Check it on a second, separate period, and keep it only if the result holds the same sign there too.
We publish every configuration we have measured this way, winners, break-evens and failures, in the measurement ledger. How the measuring works explains how to read a Strategy Tester result without fooling yourself.
Where our tools fit
GOLD DIGGA PRO runs two sleeves measured on 15-minute gold: order-block Scalps and an overnight-range breakout. A diamond appears only where a sleeve places an order, with a full trade card. GOLD DIGGA PRO ULTRA is the strategy version: the same engine as a TradingView strategy, one sleeve at a time, with an equity curve and a trade list. ZONE DIGGA shows the round-number and pivot levels on the chart, and REGIME DIGGA names the market state the regime gate is reading.