MACD-V: the MACD line measured in units of the market's own range
MACD-V is the ordinary MACD line divided by the average true range. Plain MACD is measured in price points, so a reading on ES means nothing next to one on crude oil, and a reading from a quiet year can't be compared with one from a wild year. Dividing by the range fixes that.
What it measures
Plain MACD is the gap between a fast and a slow exponential average. That gap grows when the market is volatile and shrinks when it is calm, so a fixed level means something different every year. Most traders end up reading only crosses and the zero line.
The formula in words: take the gap between the fast and the slow exponential average of the close, the same pair plain MACD uses, divide it by ATR and multiply by one hundred. ATR is the average true range, the typical size of a bar including gaps. Multiplying by one hundred only makes the number easier to read. A signal line, an exponential average of MACD-V itself, is usually drawn alongside.
The result is momentum measured in typical bar ranges.
How it behaves in real markets
Because the scale is stable, fixed levels start to mean something. High readings mark stretched moves, readings near zero mean momentum has stalled, and the same levels can be tried on ES, NQ and crude oil.
It doesn't fix the lag. MACD-V turns exactly when MACD turns; only the scale changes. The catch is low volatility: when ATR shrinks, a small move produces a large reading, and quiet holiday sessions are where that misleads.
Common ways to use it (concepts)
Fixed overbought and oversold zones
Set one upper and one lower level and use them on every market you test. Plain MACD would need different levels per contract and per year.
Momentum has stalled, stay out
When MACD-V sits between two small levels around zero, many trend rules do better if they skip the bar.
Signal-line crosses with a minimum size
Take a signal-line cross only when MACD-V is beyond a level, so tiny crosses near zero are ignored. One minimum can work across markets.
Pairing it with other indicator families
MACD-V is a momentum reading, so pair it with something that isn't. ADX tells you whether a stretched reading is the start of a run or the end of one. Bollinger Band width shows when volatility is unusually low, which is when MACD-V gets unreliable.
Stacking it with plain MACD or RSI isn't confirmation; they are all built from the same closes. I'd rather test MACD-V against plain MACD in one run and keep whichever survives out of sample.
In AlphaPrime
AlphaPrime doesn't have MACD-V as a built-in family. MACD and ATR are both built in, so the generator can already combine a MACD rule with an ATR rule, but the single normalised line is a custom indicator: ask the AI to add it and it appears under Custom indicators, ready for the strategy generator.
Test the same levels on two or three markets in one job. If one set holds up on all of them out of sample, the normalisation is doing its job. Start with the free seven-day trial.
FAQ
How is MACD-V different from plain MACD?
Same line, different scale. A reading is measured in typical bar ranges instead of price points; crosses happen on the same bars as in MACD.
Why normalise by ATR and not by price?
Dividing by price gives a percentage that still swells with volatility. Dividing by ATR removes the volatility itself.
Which levels should MACD-V use?
That is a question for testing. Try a few on several markets at once and keep the levels that hold up out of sample.