I lost a $100,000 evaluation in March to a single number I never bothered to read. EURUSD had been sliding for three days and I kept selling the bounces because my RSI was still under 30 — oversold, my notes said, so the bounce must be coming. The bounce never came. Price fell another 190 pips and my daily loss limit took the account. A week later I pulled up the same chart and added one line I had never used: the 12-period rate of change. It had crossed below zero eleven days before my first short and never looked back. My RSI was screaming oversold; the rate of change was calmly reporting that downward momentum was still accelerating. I had been trading the thermometer while ignoring the speedometer.
In this article
- The ROC formula and what each part does
- A hand-worked example on EURUSD
- Best settings for prop firm trading
- How to read ROC signals
- My worked trade: a GBPUSD long sized under my daily stop
- I tested it: ten trades with a zero-line rule
- The mistakes that cost me money
- How ROC fits prop firm rules
- Frequently asked questions
That is the cleanest way to think about this indicator. The rate of change does not tell you where price is; it tells you how fast price is getting there. Every oscillator I have covered in this series — RSI, Stochastic, Williams %R — measures momentum inside fixed walls of 0 and 100. ROC has no walls. It reports the raw percentage move over a lookback window, which makes it the most honest momentum number on my chart and also the easiest to misuse. This guide is the full treatment: what ROC measures, the formula worked by hand, the settings I use on a funded account, how to read each of its signals, a worked GBPUSD trade sized under my personal daily stop, a 10-trade test log, the mistakes that cost me money, and exactly where ROC fits inside prop firm drawdown and consistency rules.
The ROC formula and what each part does
The formula fits on one line, and I work it by hand every few months just to keep myself honest about what the indicator is really saying:
| Part of the formula | What it is | What it does to the reading |
|---|---|---|
| Close (now) | The current period’s closing price | Anchors the reading to the present — this is the “where are we” input |
| Close n periods ago | The closing price n bars back | Anchors the reading to the past — this is the “compared to what” input |
| The difference | Close now minus close n ago | Sets the sign: positive means up-momentum, negative means down-momentum |
| Divided by close n ago, times 100 | Normalised as a percentage | Makes readings comparable across pairs — a 1% move is a 1% move whether the pair is at 1.0800 or 150.00 |
Written out: ROC(n) = [(Close − Close n periods ago) ÷ Close n periods ago] × 100. The lookback n is the only setting that matters much, and it controls the indicator’s personality: a short n (9) reacts fast and lies often; a long n (25) reacts slow and rarely lies but arrives late. I run 12 as my default — a full half-day of hourly bars — and the hand-worked example below uses exactly that. The same construction is documented in markets.com’s ROC reference, which agrees on the zero-line reading and the extreme-level interpretation.
A hand-worked example on EURUSD
Real numbers, no chart needed. I am using hourly closes from a quiet Tuesday session I saved specifically because the maths was clean:
| Step | Calculation | Result |
|---|---|---|
| 1. Close 12 hours ago | The anchor candle closed at 1.0850 | 1.0850 |
| 2. Close now | The current hourly candle closed at 1.0920 | 1.0920 |
| 3. The difference | 1.0920 − 1.0850 | +0.0070 (70 pips) |
| 4. Normalise | 0.0070 ÷ 1.0850 × 100 | +0.645% |
| 5. Read it | Positive, under 1% | Mild upward momentum — price drifted, it did not surge |
That +0.645% is a typical quiet-session reading on a forex major. To calibrate your eye: on hourly charts, forex majors rarely print ROC(12) readings beyond ±2% except during news or strong trends, while gold and indices routinely hit ±4-5%. Those extreme readings are the ones that matter — a ROC(12) of +2.3% on EURUSD is not “a bit of momentum”, it is a genuine surge, and on a funded account it is exactly the kind of move you want to be positioned with rather than against. Which brings us to settings.
Best settings for prop firm trading
These are the settings I settled on after running ROC on a demo funded account for two months. They are tuned for one constraint: a daily loss limit that punishes false signals, so everything here biases toward fewer, cleaner crosses rather than more sensitive ones.
| Setting | My value | Why |
|---|---|---|
| ROC period | 12 | Half a day of hourly bars — fast enough to catch real momentum shifts, slow enough to ignore most noise |
| Signal smoothing | 3-period EMA of ROC | Takes the jagged edge off the raw line; I only act on crosses of the smoothed line, never the raw line |
| Timeframe | H1 minimum | Below H1 the zero-line crosses multiply and each one is a potential breach of my personal daily stop |
| Trend filter | 50-period EMA on price | Long signals only above the 50 EMA, short signals only below it — this one filter removed roughly half my losing crosses in testing |
| Extreme thresholds (forex) | ±2.0% | Beyond this on a major, momentum is stretched; I stop adding and start protecting |
| Extreme thresholds (gold/indices) | ±4.0% | Volatile instruments stretch further before they snap back |
| News blackout | 30 minutes each side | ROC spikes violently on red-folder news; the reading is real but untradeable, so I stand aside per my personal daily stop routine |
Two settings most guides never mention. First, the 3-period EMA of the ROC line: raw ROC(12) on H1 will cross zero 3-5 times a day in a ranging market, and each cross looks like a trade. The smoothed line cuts that to 1-2 and the ones it keeps are the ones that matter. Second, the trend filter is non-negotiable for me. An oscillator cross against the trend is a counter-trend trade wearing a costume, and counter-trend trades are how I donated that March evaluation. ROC above zero plus price above the 50 EMA is a long. ROC below zero plus price below the 50 EMA is a short. Anything else is a spectator sport.
How to read ROC signals
ROC gives four usable signals. I rank them by how much I trust each one on a funded account:
| Signal | What prints | How I use it | Trust level |
|---|---|---|---|
| Zero-line cross | Smoothed ROC crosses from negative to positive (or reverse) | Entry trigger, but only with the 50 EMA trend filter agreeing | Medium — my bread-and-butter entry |
| Extreme reading | ROC(12) beyond ±2% on a forex major | Not an entry — a warning to tighten stops and stop adding to the position | High — stretched momentum snaps back more often than it extends |
| Divergence | Price makes a higher high while ROC makes a lower high (or the bearish mirror) | Exit signal on open positions; never an entry on its own | High — divergence was the exact warning my March shorts ignored |
| Slope flattening | ROC stops rising while still positive, bars get shorter | Early warning to move the stop to breakeven — momentum is tiring | Medium — noisy, so I only act on it when I am already in profit |
Notice what is missing: there is no “overbought/oversold” row, because ROC has no fixed overbought or oversold zone the way RSI does. A +2.5% reading is extreme on EURUSD and a Tuesday on XAUUSD. If you catch yourself saying “ROC is overbought”, stop and ask overbought relative to what — the answer has to be that instrument’s own recent range, or you are borrowing someone else’s thresholds.
My worked trade: a GBPUSD long sized under my daily stop
Thursday, London session. GBPUSD had been basing above 1.2620 all morning, price was holding above the H1 50 EMA, and the smoothed ROC(12) crossed from -0.15% to +0.31% on the 09:00 candle. Trend filter green, zero-line cross fresh, no red-folder news for four hours. I took the long. Here is the sizing, built the way I build every funded-account trade — risk first, position second, checked against the max daily drawdown calculator before I clicked anything:
| Step | Number | Maths |
|---|---|---|
| Account size | $100,000 | Funded account, 5% daily loss limit = $5,000 hard ceiling |
| Personal daily stop | $2,000 | My own line, well under the firm’s — one bad day never ends the account |
| Risk on this trade | $300 (0.3%) | Conservative because this was the second trade of the day, not the first |
| Entry | 1.2680 | Market long on the zero-line cross candle close |
| Stop loss | 1.2630 (50 pips) | Below the morning base low — if that breaks, my momentum read was wrong |
| Position size | 0.6 lots | $300 ÷ (50 pips × $10 per pip) = 0.6 lots |
| Worst case | -$300 | 15% of my personal daily stop — seven consecutive losers like this still would not breach it |
| Target | 1.2780 (2R) | 2R target at the prior week’s high; partial at 1.2730 (1R) |
The trade worked: price drifted to 1.2730 by midday for the 1R partial, then caught the New York bid and tagged 1.2780 for +$600 on the remainder. Total: +$450 after the partial maths. But the number I actually cared about was the -$300. On a funded account the entry is a guess and the risk is a fact; I size the fact first and let the guess be whatever it is. ROC earned its place in my process not by predicting that Thursday, but by keeping me out of the three fake crosses earlier in the week that the 50 EMA filter vetoed.
I tested it: ten trades with a zero-line rule
To stop arguing with myself about whether ROC “works”, I ran a fixed rule on a $100,000 demo for two weeks: H1, ROC(12) smoothed with a 3 EMA, long only above the 50 EMA, short only below it, zero-line cross entries, 2R targets, $300 risk per trade (0.3%). No discretion, no news trades. Ten qualifying signals printed:
| # | Pair | Direction | Entry logic | Result (R) |
|---|---|---|---|---|
| 1 | EURUSD | Long | Zero-line cross up above 50 EMA | +2.0 |
| 2 | GBPUSD | Short | Zero-line cross down below 50 EMA | -1.0 |
| 3 | EURUSD | Long | Cross up, ROC +0.4% at entry | +1.5 |
| 4 | USDJPY | Long | Cross up above 50 EMA | +2.0 |
| 5 | GBPUSD | Long | Cross up, stopped on the retest | -1.0 |
| 6 | EURUSD | Short | Cross down, trend day | +2.0 |
| 7 | USDCHF | Short | Cross down, chopped out | -1.0 |
| 8 | AUDUSD | Long | Cross up, 1R partial then flat | +1.0 |
| 9 | USDJPY | Short | Cross down, reversed on news | -1.0 |
| 10 | EURUSD | Long | Cross up above 50 EMA | +2.0 |
Totals: 6 wins, 4 losses (60% win rate), average winner +1.83R, net +7.0R. At $300 risk per trade that is roughly +$2,100 over two weeks — about 2.1% on the account, earned without a single day threatening the daily stop. The honest footnote: trades 5, 7 and 9 were all range days where the zero-line cross was technically valid and practically a coin flip. The 50 EMA filter helps, but it cannot turn a ranging market into a trending one. On a real funded account I would have skipped the USDCHF and AUDUSD signals entirely for that reason — which is exactly the kind of discretion the test was designed to exclude, and exactly the kind I intend to keep using.
The mistakes that cost me money
Every one of these has a receipt in my journal. ROC is simple enough that the mistakes are all mine, not the indicator’s:
| Mistake | What I did | What it cost | The fix |
|---|---|---|---|
| Trading the raw line | Entered on unsmoothed ROC crosses — 4-5 signals a day in chop | Three stopped trades in one range day, -$900 | 3-period EMA of ROC; act only on the smoothed line |
| Ignoring the trend filter | Shorted a zero-line cross down while price held above the 50 EMA | -$300 and a week of tilt | No signal against the 50 EMA, ever — the March evaluation died this way |
| Borrowed thresholds | Used ±2% extremes on XAUUSD, where 4% is normal | Exited a gold runner at +$400 that ran to +$1,600 | Calibrate extremes per instrument: 2% forex, 4% gold/indices |
| News spikes | Took a +3.1% ROC surge two minutes after CPI | -$300 in 40 seconds of slippage | 30-minute blackout around red-folder news, no exceptions |
| Chasing the extreme | Bought into a +2.8% reading thinking momentum would continue | -$300 top-tick entry, the classic stretched-momentum fade | Extremes are exit/protect signals, never entries |
How ROC fits prop firm rules
An indicator is only as good as its fit with the rules that can end your account. Here is where ROC earns its keep and where it needs a chaperone:
| Prop firm rule | ROC’s fit |
|---|---|
| Daily loss limit (5%) | Excellent — the 50 EMA trend filter plus H1-minimum timeframe keeps signal count low, so daily risk stays plannable; my worst test day was -$600 |
| Maximum drawdown (10%) | Good — divergence and slope-flattening give early exit warnings before drawdown compounds, which is what the max-drawdown rule rewards |
| Consistency rule (best day cap) | Neutral — ROC finds steady 1.5-2R trends rather than lottery tickets, which naturally keeps any single day under the cap; check your consistency calculator anyway |
| News trading restrictions | Needs care — ROC spikes on news are the indicator’s least trustworthy readings; the 30-minute blackout is mandatory |
| Minimum trading days | Good — 1-2 clean signals a day on H1 fills trading-day requirements without forcing trades |
The through-line: ROC is a momentum reporter, not a strategy. Paired with a trend filter, a timeframe floor and a hard daily stop, it does one job well — it tells you which way the wind is blowing and how hard. Everything else in this guide is scaffolding to keep that one job from costing you an account.
Frequently asked questions
What is the rate of change indicator in simple terms?
It is the percentage change in price over a set number of periods. ROC(12) = +1% means price is 1% higher than it was 12 periods ago. Positive readings mean upward momentum, negative readings mean downward momentum, and the size of the number tells you how strong that momentum is.
What is the best ROC setting for day trading a prop firm account?
I use ROC(12) with a 3-period EMA smoothing on the H1 timeframe, plus a 50-period EMA trend filter on price. The smoothing cuts false zero-line crosses roughly in half, and the H1 floor keeps signal frequency low enough that a daily loss limit is never under surprise pressure.
Is ROC better than RSI for funded accounts?
Neither is better in absolute terms — they measure different things. RSI is bounded between 0 and 100 and excels at flagging overbought/oversold conditions; ROC is unbounded and excels at reporting raw momentum speed. I use RSI for reversal warnings and ROC for trend-confirmation entries. They complement rather than replace each other.
Why does ROC keep giving false signals in ranging markets?
Because ROC has no walls — in chop, price oscillates a fraction of a percent around the zero line and every wiggle is a cross. The fixes are the 3-period EMA smoothing, the 50 EMA trend filter, and simply skipping signals when the market is clearly ranging. An indicator cannot manufacture a trend that is not there.
Can I use ROC on lower timeframes like M15?
You can, but I do not on a funded account. Below H1 the zero-line crosses multiply to the point where trade frequency itself becomes the risk — more trades mean more chances to hit the daily stop. If your firm allows scalping and you want to try it, halve your risk per trade and expect the win rate from my test to drop.
Does ROC work on gold and indices or just forex?
It works on anything with a price history, but the extreme thresholds must be recalibrated — ±2% is stretched on EURUSD and ordinary on XAUUSD. Use ±4% as the starting extreme zone for gold and major indices, then adjust to each instrument’s own recent range.
Risk disclaimer: trading leveraged forex and CFD products on prop firm evaluations carries a substantial risk of loss, including the loss of challenge fees. Nothing here is financial advice — I write about what I do on my own accounts, and your rules, limits and risk tolerance are your own. Never risk money you cannot afford to lose.





3 comments
[…] open inside each bar, which is a measure of conviction inside the bar. It is also a cousin of the Rate of Change indicator, which measures how fast price moved; the RVI measures how decisively it closed. For an independent […]
[…] to clear up before we start, because this site already covers momentum from several angles. My rate-of-change article covers a single, unsmoothed ROC — KST contains four ROCs, but it is not four copies of that […]
[…] recent range, which is a different question from how persistent and unusual the last move was. The rate of change indicator measures the size of the move alone, and Connors RSI borrows that idea for only one of […]