TradingLookup

Bitcoin: are retail traders long or short?

Bitcoin: 83/100, crowded long. Small traders hold 1,177 long and 711 short contracts: 62.3% long. Week before: 66. A year before: 42. Three-year range: 1 to 85.

On 9 October 2026 small traders in CME bitcoin and micro bitcoin futures (Bybit/Binance BTCUSDT once eligible) were 62.3% long, a Crowd Gauge reading of 83 out of 100 (crowded long), according to TradingLookup's Crowd Gauge (October 2026).

The CFTC report released that day, positions as of the Tuesday before.

The contract behind the reading

Contract
CME bitcoin and micro bitcoin futures (Bybit/Binance BTCUSDT once eligible)
CFTC contract code
133741 444 weekly reports since 10 April 2018
CFTC contract code
133742 284 weekly reports since 4 May 2021
A long position is
a bet that the bitcoin price rises.
In the composite
Yes, in the crypto half

Positioning, not price. "Small traders" are everyone below the CFTC's reporting size ("non-reportable"): the closest public measure of retail positioning in CME bitcoin and micro bitcoin futures (Bybit/Binance BTCUSDT once eligible). It also holds some small firms, so treat it as a proxy.

What today's reading means

Bitcoin reads 83: small traders' long share is above about 83% of the last 156 weekly readings in this contract. The last crowded-short week (20 or less) was 28 August 2026.

Record readings for Bitcoin

Highest Crowd Gauge reading
99.6 2 weeks, first on 15 May 2020, most recently 22 May 2020
Lowest Crowd Gauge reading
0.3 on 26 November 2021
Most long small traders have been
89.3% long positions as of 19 May 2020
Least long
37.8% long positions as of 4 August 2026

Readings from 5 April 2019 (393 weeks). Long share from 10 April 2018, in the CFTC contract above with the most history.

How often Bitcoin hits an extreme

Bitcoin: weeks at an extreme, 5 April 2019 to 9 October 2026
BandWeeksShare of weeksSeparate spellsLongest spellMost recent
Crowded long (80+) 60 15% 9 16 weeks, to 19 July 2019 9 October 2026
Crowded short (20 or less) 99 25% 9 46 weeks, to 30 September 2022 28 August 2026

A new spell starts after 4 or more weeks outside the band, the same rule as the backtest's "episodes". Weeks are counted from the first published reading above.

History

Bitcoin: Crowd Gauge, last three yearsWeekly readings for Bitcoin, 0 to 100, to 2026-10-09. Latest 83.0204060801002024202520262026-10-09: 83
Last three years, weekly. Lines at 20, 40, 60 and 80 mark the bands.
Bitcoin: Crowd Gauge, full historyWeekly readings for Bitcoin from 2019-04-05 to 2026-10-09.02040608010020202021202220232024202520262026-10-09: 83
From 5 April 2019, the first week with three years of history behind it (or one year for newer contracts).

Last 12 weekly readings

Bitcoin: Crowd Gauge and CFTC small-trader positions (contract 133741), latest first
ReleasedReadingBandChangeLongShortLong share
83.0Crowded long+171,17771162.3%
66.4Leaning long+131,10980458.0%
52.9Balanced+201,28493158.0%
33.0Leaning short−341,02183255.1%
66.7Leaning long+261,27373663.4%
40.7Balanced+241,18781359.4%
16.4Crowded short+101,1211,28046.7%
6.4Crowded short+270599341.5%
4.5Crowded short+37781,08841.7%
1.3Crowded short−46521,07237.8%
5.5Crowded short−57221,06440.4%
10.6Crowded short+579386147.9%

Change: points on the 0 to 100 scale against the week before. Positions are as of the Tuesday before each release.

Did extreme Bitcoin readings come before price moves?

Average forward return in our pre-registered test (price series: FRED CBBTCUSD). 2018 onwards is the test; 2006 to 2017 is look-only.

Bitcoin: average return after extreme readings vs all weeks
Period, horizonAfter 80 or moreAfter 20 or lessAll weeksWeeks
2018 on, 1 week0.43% n=590.47% n=990.50%392
2018 on, 4 weeks9.12% n=591.11% n=992.11%390
2018 on, 12 weeks21.81% n=59-0.77% n=976.09%385

From 2018, the 4 weeks after a crowded-long Bitcoin reading averaged 9.12%, against 2.11% for all weeks; after a crowded-short reading, 1.11%. Separate crowded-long episodes in that period: 9; crowded-short: 9.

History, not a forecast. Small samples swing a lot, and one market proves little. The pooled result across markets is on the method page.

Sources for Bitcoin

Download every number: history CSV, COT observations CSV (CC BY 4.0). Spotted a mistake? Tell us.

Get our new data by email

CFD loss-rate updates, prop firm price drops and UK licence changes. Free.

Email alerts: what each list sends