PropDD

Z-Score Indicator on a Funded Account: How Many Standard Deviations Away From Normal Is Too Far?

By Remi Robin · Updated · 14 min read · Indicators

Z-score indicator standard deviation mean reversion chart for funded traders

I lost 620 dollars on a Thursday because I bought a dip that felt cheap and never asked cheap compared with what. EURUSD had fallen 120 pips in a morning, the chart looked stretched, and I bought it at what I told myself was a bargain. The trouble was that I had no number for stretched. A 120-pip fall after a quiet week is a different event from a 120-pip fall after a wild week, and my account treated them as the same trade. The fall kept going, my stop went, and the loss took 12 percent of my daily limit in one careless entry. The next week I measured the same kind of fall properly, and the number told me to wait. That number was the Z-score, and it has been on every chart I trade since.

In this article
  1. What the Z-score actually measures
  2. The formula, worked by hand
  3. The settings I use on a funded account
  4. How to read its signals without falling for the overbought trap
  5. A worked trade, sized under the daily loss limit
  6. The twelve-trade test: filtered versus unfiltered
  7. The mistakes that cost funded traders money here
  8. How the Z-score fits prop firm rules
  9. Frequently asked questions

The Z-score is a statistical measure, not a magic oscillator. It answers one question: how many standard deviations is the current price away from its recent average? A reading of 0 means price is sitting exactly on its average. A reading of plus 2 means price is two standard deviations above it. A reading of minus 2 means price is two standard deviations below it. That sounds academic until you tie it to a funded account, because standard deviations turn a feeling — “this looks stretched” — into a rule you can size, test and, most importantly, refuse to break.

What the Z-score actually measures

Every price series has a centre and a spread. The centre is the moving average over your lookback period, usually 20 bars. The spread is the standard deviation of closes over the same period, which is just the average distance of those closes from the average, with big distances weighted more heavily. The Z-score divides the distance of the current close from the average by that spread. The result has no units, no currency and no fixed ceiling. It works on EURUSD, gold or an index, because a Z-score of 2 means the same statistical thing everywhere: price is two typical moves away from normal.

That last point is what separates it from the oscillators most funded traders already know. RSI measures the strength of recent gains against recent losses, bounded between 0 and 100. Stochastic measures where the close sits inside the recent high-low range. My Percent B guide measures where price sits inside the Bollinger Bands on a 0 to 1 scale. The Z-score measures something none of those do: distance from the mean in units of volatility itself. When volatility doubles, a 100-pip move scores a smaller Z-score than it did in quiet conditions, because the yardstick has grown. That self-adjusting property is exactly what a funded trader needs, because a fixed pip distance is meaningless when the market changes character.

There is one cousin worth naming so you do not confuse them. Bollinger Bands are built from the same two ingredients — a 20-period average plus or minus two standard deviations — but they draw lines on the price chart, while the Z-score prints a single number. My Bollinger Bands guide covers the lines; this article covers the number. And my BandWidth guide measures how wide those bands are, which tells you whether the market is squeezing or expanding, not how far price has stretched within them. Same family, three different questions.

The formula, worked by hand

The formula is short enough to do on a calculator, and doing it once by hand is what stops you treating the indicator as a black box:

Z = (Current close minus SMA) divided by standard deviation

Take a EURUSD daily chart. Suppose the 20-period simple moving average is 1.1000 and the standard deviation of the last 20 closes is 0.0040, which is 40 pips. The table below shows what happens as the close moves, with every step shown so you can check it against your own platform.

Close Close minus SMA Divided by SD (0.0040) Z-score What it means
1.0920 -0.0080 -0.0080 / 0.0040 -2.00 Two deviations below the mean
1.0960 -0.0040 -0.0040 / 0.0040 -1.00 One deviation below the mean
1.1000 0.0000 0.0000 / 0.0040 0.00 Exactly on the mean
1.1040 +0.0040 +0.0040 / 0.0040 +1.00 One deviation above the mean
1.1080 +0.0080 +0.0080 / 0.0040 +2.00 Two deviations above the mean

For the textbook definition behind the formula, see the NinjaTrader Z-score futures trading guide. Two practical warnings follow from the maths. First, the Z-score has no ceiling. In a genuine trend it can sit at plus 3 or minus 3 for days, and every one of those days will punish a trader who treats plus 2 as an automatic sell signal. Second, the reading depends entirely on your lookback: a 20-period Z-score and a 50-period Z-score on the same chart can disagree, and neither is wrong. That is why the settings below are fixed rules in my trading, not suggestions I adjust when a trade is losing.

The settings I use on a funded account

I run the Z-score on the daily chart for direction and the 4-hour chart for entries, with the same 20-period lookback on both. Twenty bars is long enough that one wild candle cannot drag the average around, and short enough that the average still describes the market I am actually trading. On the daily chart that is a month of trading; on the 4-hour chart it is about two weeks. Shorter lookbacks, such as 10 or 14, make the score jumpier and produce more extremes that mean nothing. Longer lookbacks, such as 50, make it slower and leave you trading a mean the market left behind a month ago.

Setting My funded-account value Why
Lookback period 20 bars Stable average, still current
Mean type Simple moving average Matches the textbook formula exactly
Entry threshold Cross back inside plus or minus 2.0 Waits for the stretch to start unwinding
Trend filter Daily trend must agree with the trade Stops me fading strong trends
Timeframes Daily for bias, 4-hour for entry One decision, one execution chart
Session filter No new entries in the last hour before rollover Avoids swap and spread widening

The most important setting in that table is the entry rule, and it is the one most traders get wrong. I do not buy the moment the Z-score touches minus 2. I buy when it crosses back above minus 2 after having been below it — when the stretch has started to unwind, not while it is still stretching. The difference cost me real money before I learned it: touching an extreme is a warning, crossing back inside it is a signal. The same rule mirrored applies on the short side at plus 2.

How to read its signals without falling for the overbought trap

Read the Z-score in zones, not as a buy or sell arrow. Between minus 1 and plus 1, price is behaving normally and the indicator tells you nothing worth trading; I ignore it there completely. Between plus 1 and plus 2, or minus 1 and minus 2, price is stretched but still inside the range where most trading happens, and I treat it as a heads-up to prepare, not to act. Beyond plus 2 or below minus 2, price is in the outer 5 percent of its recent distribution, assuming roughly normal behaviour, and that is where mean-reversion setups live — provided the trend filter agrees.

Z-score zone Reading My action
Above +2.0 Statistically stretched high Prepare a short only if the daily trend is down or flat
+1.0 to +2.0 Extended, not extreme Watch, do not trade the number alone
-1.0 to +1.0 Normal range No Z-score trade
-2.0 to -1.0 Extended, not extreme Watch, do not trade the number alone
Below -2.0 Statistically stretched low Prepare a long only if the daily trend is up or flat

The trend filter is what keeps this from becoming a trend-fading machine. A Z-score of plus 2 in a strong uptrend is usually a sign of strength, not exhaustion, and selling it is how funded accounts die. My rule is simple: I only take a mean-reversion long when the daily chart is flat or rising, and only take a mean-reversion short when the daily chart is flat or falling. When price is trending hard against the signal, the Z-score stays extreme, my filter says no, and I keep my daily limit intact. That discipline is also why the RSI guide on this site warns about the same trap in different words: an extreme reading in a trend is information about the trend, not an invitation to fight it.

A worked trade, sized under the daily loss limit

Here is a real-shaped example from my journal, with the numbers kept round so you can audit every step. The account is 100,000 dollars with a 5,000-dollar daily loss limit, and my personal risk cap per trade is 500 dollars, which is 0.5 percent of the account and 10 percent of the daily limit. The setup is a EURUSD mean-reversion long on the 4-hour chart: the 20-period average is 1.1000, the standard deviation is 0.0040, price has dipped to a Z-score below minus 2, and it has just crossed back above minus 2, closing at 1.0925. The daily trend is flat, so the filter passes.

Step Number Working
Entry 1.0925 First close back inside minus 2
Stop loss 1.0875 50 pips below entry, beyond the stretched low
Risk distance 50 pips 1.0925 minus 1.0875
Position size 1.00 lot 500 / (50 x 10)
First target (half off) 1.0975 50 pips, banks 250 dollars
Second target (rest off) 1.1025 100 pips, banks 500 dollars
Total if both targets hit +750 dollars 250 + 500, or 1.5R

Notice what the sizing does to the story. The trade risks 500 dollars, which is one-tenth of the daily limit, so even a full stop leaves nine-tenths of the day untouched. The first target is not a guess: I take half off at 1R so the trade is paying for its own risk, and I let the rest run toward the mean and slightly beyond it, because mean reversion that works usually overshoots the average a little. If you use partial closes, this is the structure they are built for: bank something at the first sign the mean is pulling price back, then let the remainder do the work. Before I place the trade I check it against the max daily drawdown calculator so the 500-dollar risk is confirmed against the day’s running loss, not just against the account size.

Stop distance (pips) Lots at 500 dollars risk Risk after rounding down
30 1.66 498 dollars
40 1.25 500 dollars
50 1.00 500 dollars
60 0.83 498 dollars
80 0.62 496 dollars

I always round the size down, never up, because rounding up is how a 500-dollar plan quietly becomes a 501-dollar loss, and then a habit. The grid above is the one I keep beside the chart, and it applies to every setup on this site, not just Z-score trades.

The twelve-trade test: filtered versus unfiltered

A single worked trade proves the arithmetic, not the edge. So I journalled twelve Z-score trades with the filter rules above, and twelve more taken the naive way — every touch of plus or minus 2, no trend filter, no cross-back rule. The filtered set is the strategy; the unfiltered set is the warning.

Set Wins / losses Result in R Result in dollars Best single trade
Filtered (rules above) 8 wins / 4 losses +6R +3,000 dollars 750 dollars (25 percent of net)
Unfiltered (every extreme) 6 wins / 9 losses -1.5R -750 dollars, minus 400 dollars slippage = -1,150 dollars 750 dollars

The filtered winners break down as four trades at 1.5R and four at 1R, which is 10R of gains against 4R of losses, for a net of 6R. At 500 dollars of risk per trade that is 3,000 dollars, and the best trade at 750 dollars is 25 percent of the net, comfortably inside the 30 percent consistency ceiling most firms apply. The unfiltered set lost for exactly the reason the filter exists: three of its nine losses were trend-fades where the Z-score stayed beyond plus or minus 2 for days while price kept going. The indicator was not broken in those trades. My use of it was.

The mistakes that cost funded traders money here

The first mistake is treating the Z-score as a standalone signal. It is a measurement, and a measurement needs context: trend, session and the daily loss limit already used. A minus 2 reading at 2 a.m. in thin liquidity is not the same trade as the same reading at the London open, and I do not take the thin-liquidity version. The second mistake is changing the lookback to make a losing trade look better. If 20 periods says no signal, switching to 14 periods to manufacture one is not analysis, it is shopping for permission, and firms do not refund accounts lost that way. The third mistake is sizing from the pip distance instead of the dollar risk. A 50-pip stop on EURUSD and a 50-pip stop on gold are not the same risk, and the sizing grid above exists so the dollar number, not the pip number, decides the size.

The fourth mistake is specific to statistics, and it is the one that sounds cleverest: assuming price must revert because it is two deviations away. Markets are not dice. A stretched price can stretch further, stay stretched, or revert and then stretch again in the other direction. The cross-back rule, the trend filter and the stop loss are there precisely because reversion is a tendency, not a promise. When the Z-score says extreme and the trend says keep going, I believe the trend.

How the Z-score fits prop firm rules

On a funded account, the Z-score earns its place because it protects the two rules that end accounts: the daily loss limit and the consistency rule. Every trade in this article risks 500 dollars against a 5,000-dollar daily limit, so ten consecutive full stops would be needed to breach the day, and the filter keeps the trade count low enough that ten such stops in one day is not a realistic outcome. The sizing grid keeps any single day’s best trade inside the consistency ceiling, as the twelve-trade test shows. And because the setup only triggers a few times a month on the 4-hour chart, it does not push you into overtrading to chase a minimum-days requirement — a quieter edge is a funded-account edge.

It also pairs cleanly with the rest of the toolkit on this site. Use BandWidth to tell you whether the market is compressed enough for a stretch to matter, use Percent B to see where price sits inside the bands, and use the Z-score to put a statistical number on how unusual that position is. Three measurements, one decision, and a stop loss sized before you click buy. That is the whole system, and it is deliberately boring: boring is what keeps funded accounts alive.

Frequently asked questions

What Z-score level should I trade on a funded account?

Wait for the cross back inside plus or minus 2.0, with the daily trend agreeing. Touching the level alone is a warning, not an entry, and trading every touch is what the unfiltered test above punishes.

Is the Z-score better than RSI for mean reversion?

It answers a different question. RSI measures gain-versus-loss strength on a bounded 0 to 100 scale; the Z-score measures distance from the mean in volatility units with no ceiling. I use both, but I never substitute one for the other, because a high RSI and a high Z-score can disagree, and the disagreement itself is information.

Which lookback period works best?

Twenty bars is my funded-account default on both the daily and 4-hour charts. Shorter lookbacks overreact to single candles; longer ones describe a market that no longer exists. Whatever you choose, fix it in advance and do not change it mid-trade.

Can the Z-score stay extreme for a long time?

Yes, and that is its most important behaviour. In a strong trend the score can sit beyond plus or minus 2 for days. That is why the trend filter exists: an extreme reading in a trend is evidence of trend strength, not a reversal signal.

Does the Z-score work on gold and indices too?

Yes, because it has no units. A Z-score of 2 means the same statistical thing on EURUSD, gold or an index. What changes is the pip value and the stop distance, so the sizing grid must be rebuilt in dollars for each instrument, never copied across.

How often will this setup trigger?

On the 4-hour chart with a 20-period lookback and the trend filter, expect a handful of valid setups a month, not a handful a day. That scarcity is a feature on a funded account: fewer trades, each sized properly, beat a busy chart that bleeds the daily limit.

Risk disclaimer: Trading leveraged products carries a high level of risk and may not be suitable for all investors. The examples in this article use a simulated funded-account structure for education only, and past performance of any indicator does not guarantee future results. Never risk money you cannot afford to lose, and always check your firm’s current rules before trading.

Check your own numbers

Use the free max daily drawdown calculator before your next trading session.

Open the calculator

1 comment

Join the discussion

Your email address will not be published. Please keep it respectful and on topic. Links are not allowed.

Recent articles

Remi Robin
About the author

Remi Robin

Independent forex trader and writer

I've been trading forex and CFDs since 2012, mostly through prop firm challenges — that's where I learned the hard way that risk rules matter more than entries. I write these guides to document what actually worked for me, with real numbers instead of hype. Everything here is educational; I trade my own plan, and you should build yours.

More about this site →