Every 3-bar pivot on 18 years of futures data — indices, metals, energy and FX,
from the 5-minute chart (456,231 pivots on Nasdaq alone) up to 1-hour —
tracked from the moment it printed until the moment it was taken out (or the day
closed). The heat map shows what a pivot at each hour, on each day of the week, was
actually worth.
Pick a market — everything below follows your choice: the heat map, the how-to
chart, the timeframe stack and the takeaways.
Pivot run distance — hour of day × day of week
8 futures markets · 2008–2025 · every 3-bar pivot from the
5-minute to the 1-hour chart, each tracked to its break or the day's end.
Oil excludes mid-March to mid-May 2020, when its price series traded near or below zero.
The metric. When a pivot prints, we measure how far price travels
away from it — pivot high to the lowest low (or pivot low to the highest high)
— before the pivot is taken out or the trading day ends, whichever comes first.
The run is the gap between the two dashed lines, shown as % of the pivot
price. Pivot lows mirror this upside down.
How to use it — run levels on a (made-up) chart
— its own timeframe, pivots and hour; the market follows the buttons above
When the pivot candle opened (ET), not when the pivot confirmed
The 25th line is the nearest level: three in four
pivots ran further than it before they were taken out or the day ended.
Stack the timeframes: run targets that grow as the pivot confirms
How far have pivots like this run before they were taken out, and how does that grow as higher timeframes confirm? That’s
where stacking the timeframes comes in.
Think of it like dominoes
A higher-timeframe pivot can’t exist before the lower-timeframe one. The push that makes a
15-minute pivot has to start as a 5-minute pivot, and the 5-minute chart confirms it first: three 5-minute candles
close long before three 15-minute ones do. Then comes the 30-minute, then the hourly. Each timeframe is a domino
that can only fall after the one before it, and each fallen domino brings that timeframe’s run levels into play.
If you are in before the 5-minute pivot has even confirmed, the 5-minute run levels are the first,
most aggressive targets. The nearest of them, the 25th percentile, is the high-odds first target: three in
four pivots ran further. If price keeps going and a 15-minute pivot confirms, you know more, and the 15-minute
levels come into play; then the 30-minute, and so on. That first target steps out with every domino that falls.
The targets grow as the chain grows, which gives a statistical reason to hold part
of a position for the larger move.
Pivot confirmed onPivot
Pivot low and pivot high show the same levels: the study’s all-hours summary per timeframe pools highs and lows.
The failure branch
Run levels by timeframe · · all hours
Pivot
Pivots
25th
Median
75th
85th
Never out
Run from the pivot’s extreme until it is taken out or the day ends, as % of price,
2008–2025, highs and lows together. “25th” = three in four pivots ran further than this.
“Never out” = share of pivots not taken out before the day ended.
How to read these numbers. Each timeframe’s levels come from that
timeframe’s own pivots, measured from the pivot’s extreme price, exactly like the heat map above. They are
not conditional on the smaller pivot having confirmed first, and by the time a higher-timeframe pivot confirms, part
of its run may already be behind you. Treat them as the typical size of each timeframe’s move, not as a
forecast for this one.
Educational research, not financial advice. These are historical distributions from futures
data; they describe what happened, not what will happen. Place and size your own stops and targets for your own
risk.
Key takeaways
The open pays, the close doesn't. A Nasdaq pivot
printed in the 09:00–10:00 block runs a median 0.22% before it's taken
out or the day ends — about 3.4× the 0.07% you get
from a 16:00 pivot.
The fattest cell: Friday 09:00.
Those pivots average 0.43% of travel — the biggest average in
Nasdaq's map, versus 0.08% for the thinnest cell
(Friday 16:00).
Median ≠ mean — mind the tail. Across all
456,231 of Nasdaq's 5-minute pivots the median run is 0.10% but the
mean is 0.23%. The average is flattered by a minority of runners; the
median is the honest "what usually happens".
Know your overnight. The 03:00 London hour runs a
0.11% median on Nasdaq — and 13% of all 5-minute
pivots are never taken out before the day ends at all.
Scale the pivot, scale the run. Switch the timeframe
filter: a 1-hour Nasdaq pivot printed at 09:00 runs a median 0.65%
before it breaks — 3.0× the 0.22% of a 5-minute
pivot at the same hour — and 39% of 1h pivots never break
at all (versus 13% on 5-minute).
The open pays, the close doesn't. An S&P pivot
printed in the 09:00–10:00 block runs a median 0.24% before it's taken
out or the day ends — about 2.9× the 0.08% you get
from a 16:00 pivot.
The fattest cell: Friday 09:00.
Those pivots average 0.45% of travel — the biggest average in
the S&P's map, versus 0.10% for the thinnest cell
(Friday 16:00).
Median ≠ mean — mind the tail. Across all
349,410 of the S&P's 5-minute pivots the median run is 0.14% but the
mean is 0.29%. The average is flattered by a minority of runners; the
median is the honest "what usually happens".
Know your overnight. The 03:00 London hour runs a
0.15% median on the S&P — and 15% of all 5-minute
pivots are never taken out before the day ends at all.
Scale the pivot, scale the run. Switch the timeframe
filter: a 1-hour S&P pivot printed at 09:00 runs a median 0.66%
before it breaks — 2.8× the 0.23% of a 5-minute
pivot at the same hour — and 41% of 1h pivots never break
at all (versus 15% on 5-minute).
The open pays, the close doesn't. A Dow pivot
printed in the 09:00–10:00 block runs a median 0.23% before it's taken
out or the day ends — about 3.1× the 0.07% you get
from a 16:00 pivot.
The fattest cell: Thursday 09:00.
Those pivots average 0.45% of travel — the biggest average in
the Dow's map, versus 0.09% for the thinnest cell
(Friday 16:00).
Median ≠ mean — mind the tail. Across all
440,298 of the Dow's 5-minute pivots the median run is 0.12% but the
mean is 0.26%. The average is flattered by a minority of runners; the
median is the honest "what usually happens".
Know your overnight. The 03:00 London hour runs a
0.14% median on the Dow — and 13% of all 5-minute
pivots are never taken out before the day ends at all.
Scale the pivot, scale the run. Switch the timeframe
filter: a 1-hour Dow pivot printed at 09:00 runs a median 0.66%
before it breaks — 2.9× the 0.23% of a 5-minute
pivot at the same hour — and 40% of 1h pivots never break
at all (versus 13% on 5-minute).
The open pays, the close doesn't. A gold pivot
printed in the 09:00–10:00 block runs a median 0.21% before it's taken
out or the day ends — about 3.5× the 0.06% you get
from a 16:00 pivot.
The fattest cell: Friday 08:00.
Those pivots average 0.39% of travel — the biggest average in
gold's map, versus 0.07% for the thinnest cell
(Tuesday 16:00).
Median ≠ mean — mind the tail. Across all
444,563 of gold's 5-minute pivots the median run is 0.11% but the
mean is 0.23%. The average is flattered by a minority of runners; the
median is the honest "what usually happens".
Know your overnight. The 03:00 London hour runs a
0.13% median on gold — and 16% of all 5-minute
pivots are never taken out before the day ends at all.
Scale the pivot, scale the run. Switch the timeframe
filter: a 1-hour gold pivot printed at 09:00 runs a median 0.55%
before it breaks — 2.6× the 0.21% of a 5-minute
pivot at the same hour — and 47% of 1h pivots never break
at all (versus 16% on 5-minute).
The open pays, the close doesn't. An oil pivot
printed in the 09:00–10:00 block runs a median 0.69% before it's taken
out or the day ends — about 3.8× the 0.18% you get
from a 16:00 pivot.
The fattest cell: Wednesday 10:00.
Those pivots average 1.41% of travel — the biggest average in
oil's map, versus 0.25% for the thinnest cell
(Monday 16:00).
Median ≠ mean — mind the tail. Across all
428,434 of oil's 5-minute pivots the median run is 0.36% but the
mean is 0.81%. The average is flattered by a minority of runners; the
median is the honest "what usually happens".
Know your overnight. The 03:00 London hour runs a
0.40% median on oil — and 15% of all 5-minute
pivots are never taken out before the day ends at all.
Scale the pivot, scale the run. Switch the timeframe
filter: a 1-hour oil pivot printed at 09:00 runs a median 2.03%
before it breaks — 3.0× the 0.69% of a 5-minute
pivot at the same hour — and 43% of 1h pivots never break
at all (versus 15% on 5-minute).
The open pays, the close doesn't. A pound pivot
printed in the 09:00–10:00 block runs a median 0.14% before it's taken
out or the day ends — about 3.0× the 0.05% you get
from a 16:00 pivot.
The fattest cell: Friday 08:00.
Those pivots average 0.25% of travel — the biggest average in
the pound's map, versus 0.06% for the thinnest cell
(Monday 16:00).
Median ≠ mean — mind the tail. Across all
393,872 of the pound's 5-minute pivots the median run is 0.09% but the
mean is 0.17%. The average is flattered by a minority of runners; the
median is the honest "what usually happens".
Know your overnight. The 03:00 London hour runs a
0.13% median on the pound — and 16% of all 5-minute
pivots are never taken out before the day ends at all.
Scale the pivot, scale the run. Switch the timeframe
filter: a 1-hour pound pivot printed at 09:00 runs a median 0.37%
before it breaks — 2.8× the 0.13% of a 5-minute
pivot at the same hour — and 47% of 1h pivots never break
at all (versus 16% on 5-minute).
The open pays, the close doesn't. A euro pivot
printed in the 09:00–10:00 block runs a median 0.13% before it's taken
out or the day ends — about 3.2× the 0.04% you get
from a 16:00 pivot.
The fattest cell: Thursday 08:00.
Those pivots average 0.25% of travel — the biggest average in
the euro's map, versus 0.05% for the thinnest cell
(Monday 16:00).
Median ≠ mean — mind the tail. Across all
372,771 of the euro's 5-minute pivots the median run is 0.08% but the
mean is 0.15%. The average is flattered by a minority of runners; the
median is the honest "what usually happens".
Know your overnight. The 03:00 London hour runs a
0.11% median on the euro — and 17% of all 5-minute
pivots are never taken out before the day ends at all.
Scale the pivot, scale the run. Switch the timeframe
filter: a 1-hour euro pivot printed at 09:00 runs a median 0.35%
before it breaks — 2.8× the 0.12% of a 5-minute
pivot at the same hour — and 47% of 1h pivots never break
at all (versus 17% on 5-minute).
The open pays, the close doesn't. An Aussie pivot
printed in the 09:00–10:00 block runs a median 0.18% before it's taken
out or the day ends — about 2.7× the 0.07% you get
from a 16:00 pivot.
The fattest cell: Friday 08:00.
Those pivots average 0.37% of travel — the biggest average in
the Aussie's map, versus 0.09% for the thinnest cell
(Monday 16:00).
Median ≠ mean — mind the tail. Across all
377,438 of the Aussie's 5-minute pivots the median run is 0.13% but the
mean is 0.26%. The average is flattered by a minority of runners; the
median is the honest "what usually happens".
Know your overnight. The 03:00 London hour runs a
0.15% median on the Aussie — and 17% of all 5-minute
pivots are never taken out before the day ends at all.
Scale the pivot, scale the run. Switch the timeframe
filter: a 1-hour Aussie pivot printed at 09:00 runs a median 0.48%
before it breaks — 2.7× the 0.18% of a 5-minute
pivot at the same hour — and 48% of 1h pivots never break
at all (versus 17% on 5-minute).