PropDD

McGinley Dynamic: The Adaptive Average That Ended My Chop-Week Bleed

By Remi Robin · Updated · 14 min read · Indicators

McGinley Dynamic self-correcting average chop filter chart for funded traders

March gave me a week I still think about whenever a chart goes sideways. My funded account was up, the trend on the daily chart was clearly higher, and my moving average kept telling me to buy every dip. I took nine trades in four days, and seven of them stopped out. None of the losses was large on its own, but together they ate almost half my daily loss limit twice over, and the account never actually went anywhere. The average was not wrong about the trend. It was wrong about the noise, and on a funded account noise is not free. Every whipsaw costs a slice of a limit you cannot buy back.

In this article
  1. What the line is actually measuring
  2. The self-correcting formula, worked by hand
  3. Settings that fit a funded account
  4. How I read its signals
  5. Sizing a McGinley trade under a daily loss limit
  6. I tested it: ten trades taken, two skipped
  7. Where the line fails
  8. Frequently asked questions

That week pushed me to an indicator I had ignored for years because its name sounded like a gimmick. The McGinley Dynamic, built by market technician John McGinley, is a moving average that corrects itself for the speed of price. When price moves fast, the line adjusts its own smoothing so it does not chase every tick. When price slows down, it tightens up and tracks more closely. I replaced my old average with it as a trend filter, not as a signal machine, and the chop-week bleed stopped. Not because the line predicts anything, but because it stopped giving me permission to trade noise.

This article explains what the McGinley Dynamic measures, its formula worked by hand with real numbers, the settings that fit a funded account, how I read its signals, and the sizing maths that keep a filtered trade inside a daily loss limit. If your average keeps flipping you in and out of sideways markets, this is the filter I wish I had fitted before that March week.

What the line is actually measuring

A normal moving average answers one question: where has price been on average over the last few bars? The McGinley Dynamic answers a slightly different one: where is price on average, adjusted for how fast it got there? That second part matters. A fast spike and a slow drift can end at the same price, but they do not mean the same thing. The spike is often exhaustion. The drift is often a real trend forming. A plain average treats both the same. The McGinley line does not.

The trick is a denominator that changes with every bar. When the latest close is far above the previous line value, the ratio of close to line grows, the denominator grows, and the line moves a smaller step than it otherwise would. When the latest close drops sharply below the line, the ratio shrinks, the denominator shrinks, and the line steps down faster relative to its own baseline. In plain language, the line resists being dragged by sudden jumps and catches down-moves sooner than a fixed average would. That is the whole point of the design: less lag than a simple average in a trend, less whipsaw than an exponential average in chop.

It is still a moving average, so it still lags. It will not catch a top, it will not warn you before a news spike, and on its own it is not a strategy. I use it the way I use every indicator on this site: as a filter that decides which of my setups I am allowed to take, and which ones I must leave alone. The setups themselves still come from structure, and the risk still comes from the daily limit.

The self-correcting formula, worked by hand

The formula looks intimidating until you run one bar through it. Let MD stand for the previous McGinley value, C for the latest close, and k for a constant equal to 0.6 times your chosen period. For a 10-period line, k equals 6. The new value is:

Step Calculation Result
1. Ratio of close to previous line 1.0900 / 1.0850 1.00461
2. Raise the ratio to the fourth power 1.00461 ^ 4 1.01856
3. Adaptive denominator 6 x 1.01856 6.1114
4. Price distance to cover 1.0900 – 1.0850 0.0050
5. Adaptive step 0.0050 / 6.1114 0.00082
6. New McGinley value 1.0850 + 0.00082 1.0858

Notice what happened in step 5. Without the speed adjustment, the step would have been 0.0050 divided by 6, which is 0.00083. The adjustment barely changed it on this quiet bar, because price had not moved fast. Now run a sharp fall through the same maths. Previous line 1.0850, close 1.0700: the ratio is 0.98618, raised to the fourth power it is 0.94578, the denominator shrinks to 5.6747, and the step becomes -0.0150 divided by 5.6747, which is -0.00264. The unadjusted step would have been -0.00250. So on the falling bar the line steps down faster than its baseline, and on the rising bar it steps up slower than its baseline. That asymmetry is deliberate. McGinley wanted the line to ride rising markets patiently and react to falling markets promptly, and the fourth power is what makes the adjustment strong enough to matter without making the line jump around.

Two practical consequences follow. First, the line needs a seed value, usually the first close, and a few bars to settle, so do not judge it on the first five bars of a fresh chart. Second, the period you choose changes k, and therefore changes everything. A 10-period line with k equal to 6 reacts roughly twice as fast as a 20-period line with k equal to 12. Pick the period for the job, not for the round number.

Settings that fit a funded account

I tested three periods on the hourly and four-hour charts before settling. Faster is not better here: more signals in chop means more slices off the daily limit. The funded-account question is never which setting catches the most moves. It is which setting keeps you out of the most noise while still catching the moves that pay.

Setting What it does Funded-account verdict
McGinley 10, H1 chart Responsive, tracks intraday swings closely My default for entries; needs a structure filter beside it
McGinley 20, H4 chart Smooth, ignores most intraday chop My trend filter; I only take H1 longs when H4 line slopes up
McGinley 10, M15 chart Very responsive, flips often in ranges Rejected; produced the same whipsaw the line was meant to fix
McGinley 30, daily chart Very slow, rarely changes direction Useful for weekly bias only, too slow for challenge pacing

My working combination is simple: the H4 McGinley 20 sets the direction, and the H1 McGinley 10 times the entry. If the H4 line slopes down, I do not take H1 longs no matter how clean they look. That rule alone would have saved five of my seven March whipsaw losses. The QuantConnect McGinley Dynamic reference uses a 10-period default, which matches my H1 choice, and I keep the H4 line at 20 because a funded account rewards patience more than speed.

How I read its signals

I read exactly three things from the line, and I ignore everything else. First, slope: a rising line means the recent speed-adjusted trend is up, a falling line means it is down, and a flat line means chop, which means no new trades. Second, position: price above a rising line supports longs, price below a falling line supports shorts, and price crossing back and forth across a flat line is noise, not opportunity. Third, distance: when price stretches far above a rising line after a fast run, I do not chase. I wait for price to return toward the line, because the line itself tells me the fast move is the least reliable part of the trend.

What I do not do is trade every crossover. A crossover of price and the McGinley line in a flat market is the same trap as a crossover of any two averages: it fires, reverses, and fires again. My rule is that a crossover only counts if the higher-timeframe line agrees with the direction and price has already formed a structure I recognise, such as a pullback to a prior swing or a break of a small range. The line is the permission slip. Structure is the trade.

One more reading habit matters. When the line flattens after a trend, I treat it as a tired-trend warning, not a reversal signal: I tighten stops and stop adding to winners, but I do not flip short on a flat line alone.

Sizing a McGinley trade under a daily loss limit

The filter is useless if the size is wrong, because a filtered trade at double size still breaches the same limit. My account is 100,000 dollars, the firm daily loss limit is 5,000 dollars, and my personal daily stop is 2,500 dollars, half the firm limit. My base risk is 500 dollars per trade, which is 0.5 percent of the account and one fifth of my personal daily stop. Five base-risk losses in a day end my day. That is the point.

Sizing step Worked example Number
Entry after close above rising H1 line EURUSD long at 1.0885 Entry 1.0885
Stop below the line and the swing low Stop at 1.0845, below McGinley 1.0858 40 pips of risk
Lots from base risk 500 dollars / (40 pips x 10 dollars per pip per lot) 1.25 lots
First target, one times risk 1.0925 +500 dollars
Second target, two times risk 1.0965 +1,000 dollars
Share of personal daily stop if stopped out 500 / 2,500 dollars 20 percent

If the stop were 80 pips instead of 40, the same 500 dollars would buy only 0.63 lots: the line sets the stop, the stop sets the size, never the other way round. Run your own numbers through the max daily drawdown calculator before the session, not after a loss, because after a loss the temptation is to size up, and sizing up after a loss is how a filtered strategy still breaches an account.

I tested it: ten trades taken, two skipped

For two weeks I traded only setups that passed both filters: H4 McGinley 20 sloping in the trade direction, and an H1 entry above a rising line for longs or below a falling line for shorts. Everything else went in the journal as a skipped trade, not as a missed opportunity. The log below is the full set, wins and losses together, because a filtered log that hides its losers is marketing, not testing.

Trade Setup Result R multiple
1 EURUSD long, H1 pullback to line Win +1.8R
2 GBPUSD long, H4 slope up, H1 break Win +2.2R
3 EURUSD long, flat H4 line ignored rule, skipped Skipped, not traded 0R
4 EURUSD short, H4 slope down Loss -1.0R
5 XAUUSD long, return to rising line Win +1.5R
6 GBPUSD short, H1 cross, H4 flat, skipped Skipped, not traded 0R
7 EURUSD long, H1 pullback to line Win +2.0R
8 XAUUSD short, H4 slope down Loss -1.0R
9 EURUSD long, trend continuation Win +1.2R
10 GBPUSD long, H1 break with H4 slope Win +2.4R
11 EURUSD short, late entry after flat line Loss -1.0R
12 XAUUSD long, return to rising line Loss -1.0R

Counted honestly, the traded set is six wins and four losses. The winners total 11.1R and the losers total 4.0R, for a net of 7.1R. At 500 dollars of risk per trade, that is 3,550 dollars across the traded set, and no single day came close to my personal daily stop, because the worst day held two losses totalling 1,000 dollars. The two skipped trades matter as much as the winners: both were the exact kind of flat-line crossover that cost me money in March, and skipping them cost nothing. A filter that only ever says yes is not a filter.

For comparison, the EMA guide on this site shows the same pullback logic with a fixed average. The practical difference was not entry price but entry count: fewer trades meant fewer chances to bleed the daily limit in chop.

Where the line fails

The line fails in exactly the places every average fails, and pretending otherwise would be dishonest. In a sharp news spike it lags, then overreacts, then flattens, and all three phases can look like signals if you want them to. I do not trade the line within thirty minutes of major news, and no filter changes that rule. In a long, slow range it flattens and stays flat, which is actually the line working correctly, but traders who need action will still trade the flat line and still lose. The failure there is not the indicator. It is the impatience.

Mistake What happens The fix I use
Trading every crossover Repeated whipsaw in ranges Require H4 slope agreement plus structure
Using M15 as the main line Too many flips, too many fees H1 for entries, H4 for direction, nothing faster
Chasing price far from the line Buying exhaustion, stopping out on the return Wait for the return toward the line before entering
Flipping on a flat line Reversal trades with no trend behind them Flat line means no new trades, not a reversal
Sizing from the entry, not the stop One loss eats a quarter of the daily stop Size from stop distance, capped at 0.5 percent risk

It also fails as a standalone system in trending markets that move in steps rather than smoothly. A stepped trend, where price jumps, consolidates, then jumps again, will pull the line up in bursts and leave it behind during consolidations. Entries taken during the consolidation, expecting the line to keep rising, can stop out on the next small dip. The fix is the same as everywhere else in this article: the line filters, structure decides, and the daily limit has the final say.

Related reading on this site: the trend pullback plan pairs naturally with a rising McGinley line, the choppiness filter covers the flat-line problem from a different angle, and the give-back article explains why a quiet, filtered week beats a busy, profitable-looking one that ends red. Together they are the same lesson from four sides: on a funded account, the trades you skip protect the trades you take.

Frequently asked questions

Is the McGinley Dynamic better than an EMA?

It is different, not universally better. It adapts its smoothing to price speed, so it whipsaws less in chop and tracks trends with less lag than a simple average. In a clean trend, a well-chosen EMA will often give a similar entry. The McGinley advantage shows up in sideways markets, where fewer false signals protect the daily loss limit.

What period should a funded trader start with?

Start with 10 on the hourly chart for entries and 20 on the four-hour chart for direction. That combination is responsive enough to trade and slow enough to filter. Do not start on the fifteen-minute chart, where the line flips often enough to recreate the whipsaw problem it solves.

Can I trade McGinley crossovers alone?

You can, but on a funded account you probably should not. Crossovers in flat markets are the line’s weakest signal. Require agreement from the higher-timeframe line and a recognisable structure, such as a pullback or a small range break, before you risk capital on a crossover.

Does the line work on gold and indices as well as forex?

Yes, with one adjustment: gold moves faster and gaps more, so the same 40-pip stop logic does not transfer directly. Size from the actual stop distance in dollars, not from a pip count borrowed from a forex pair, and expect wider stops on gold to mean smaller size for the same risk.

Why does the line need a few bars to settle?

The formula is recursive: each new value depends on the previous one. The first value is usually seeded with the first close, so early values carry that seed’s influence. After ten to twenty bars the seed’s effect is negligible, which is why you should never judge the line on a freshly opened chart.

What is the single biggest mistake with this indicator?

Treating a flat line as a signal. A flat McGinley line means the market is undecided, and undecided markets punish action. The profitable move on a flat line is usually no move at all, and the discipline to accept that is worth more than any setting tweak.

Risk disclaimer: This article is for education only and is not financial advice. Trading leveraged products on a funded account can result in the loss of the account. Test any indicator on a demo account, respect your firm’s rules, and never risk money you cannot afford to lose.

Check your own numbers

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

Open the calculator

3 comments

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 →