Fair Value Gap vs Supply and Demand
Michael Whitman
·
Trading since 2017, full-time since 2020
·
YouTube creator, 137K subscribers
·
Published 2026-09-08
·
Updated
·
Reviewed
A fair value gap is the strip between the outer wicks of three candles, so the rule sets its width. A supply or demand zone is drawn around the quiet area a move left from, usually a candle or more. On TLT the gap was narrower, and width shaped how often it filled.
Supply and demand zones and fair value gaps mark the same kind of place: where price left fast. The practical difference is width. A zone is drawn around the candle the move started from; a gap is only the space the move skipped. Measured in average ranges on the long-bond fund TLT, that width decided a lot about what happened next.
What each one is
The gap’s width comes from two wicks. Whatever price range the wicks of candles one and three leave open is the gap, and the fair value gap page gives the rule used here.
A supply or demand zone is the quiet area price dropped or rallied away from, drawn by hand around the base candle or candles. Supply and demand explains how the zone is usually drawn.
Both mark a departure. The zone marks where it began, rather than the stretch it skipped, which is what the gap marks. On a bullish move the zone sits below the gap, since the base candle is candle one.
Where they differ
Who decides the edges. The gap’s edges are two wicks, the same for everyone, whereas a zone’s edges depend on which candles the trader includes and whether wicks or bodies are used.
How wide each is. On TLT, candle one’s own range was a median 0.693 of a day’s average true range, while the gap above it was a median 0.459. A zone of one candle or more is the wider area.
What a touch means. A zone is wide enough that price often trades inside it without going through, while a narrow gap is often crossed in a single session.
How they can be counted. Gaps can be listed by a script. Zones need the drawing rule written first, which is why this page measures the gap and uses candle one as the zone’s minimum width.
Which one to use: Fair Value Gap or Supply and Demand?
Use the supply or demand zone when your stop has to survive ordinary noise, since the extra width is room for price to move without voiding the idea.
Prefer the gap when the edges must be identical for everyone, and look at its width in average ranges before treating it as an area at all.
When a gap is tiny relative to the day’s range, treat it as noise. On TLT almost all of those were traded straight through within a month.
Does gap width change the fill?
The test. For every gap on TLT with 20 sessions after it, divide the band’s width by the 14-day Wilder average true range as of candle one. Then count, over the 20 sessions after candle three, which gaps price re-entered and which it crossed completely.
Re-entry barely moved with size. In each of four size groups, between 86.25% and 90.26% of gaps were re-entered within 20 sessions. The full fill did move.
Full fills by width: under 0.10 range, 285 of 327 (87.16%); 0.10 to 0.25, 398 of 462 (86.15%); 0.25 to 0.50, 452 of 582 (77.66%); 0.50 or more, 774 of 1,177 (65.76%). Smallest against largest, the difference is too large to put down to chance (Fisher exact test, p below 0.001).
An August 2026 bond gap
The candles. On 21 August 2026 TLT’s high was $82.27 and its low $81.90. On 25 August the low was $82.97. The band from $82.27 to $82.97, 70 cents wide, was a bullish gap, while candle one, the base a demand zone would be drawn around, spanned 37 cents.
In average ranges. The 14-day average true range on 21 August was $0.662, so the gap measured 1.058 ranges and candle one 0.559. This was one of the wide gaps.
What followed. On 28 August the low of $82.78 came back into the band, and on 31 August the low of $82.15 went through its far edge. The gap was filled four sessions after candle three; the demand zone under it was traded into the same day.
The original data
The sample. 6,084 TLT daily bars from Yahoo Finance, 30 July 2002 to 2 October 2026, the high, low and close of each session. Of all the gaps, 2,548 had an average range at candle one and 20 sessions after.
Across all of them. 88.42% were re-entered and about three in four fully filled within 20 sessions. The widest was the March 2020 bullish gap from $158.82 to $170.34, 4.67 ranges, fully filled two sessions after it formed. In 1,726 of the 2,548 gaps the band was narrower than candle one’s range.
The file. Every gap with its width in ranges, candle one’s range and both 20-session outcomes is in the TLT gap size file.
What is left out. No zone is drawn by hand, so candle one’s range stands in as the narrowest zone a trader would mark. What happened after a fill is not measured.
When it fails
The failure is trading a gap as if width did not matter. A gap a tenth of a day’s range wide is crossed by an ordinary session, so a plan that buys at its top edge with a stop under its bottom edge is a plan with a stop inside the noise. On TLT those slivers were filled through almost nine times in ten within a month, which is just what noise looks like.
The second failure is widening a zone after price trades through it. The area then never fails.
A third is comparing gap widths in dollars across years. Measure them in ranges instead.
A fourth is treating the re-entry as a hold. Nearly every gap was re-entered.
And a fifth is drawing the zone after the move. Its edges then fit the reaction.
Related
Fair value gap has the band rule. Supply and demand sets out how traders draw the zones. And imbalance is the other name this site gives the same three-candle gap.
Measure a gap in average true ranges before you plan around it. A sliver of a gap on the bond fund was filled almost every time, which makes it a poor area to expect a reaction from. A wide one was left partly open more often, so a stop inside it gets tested less.
— Michael Whitman
This page is educational, not financial advice. Test every idea on your own charts before risking money.