How many observations does gold seasonality have?

11 or 12 per calendar month, using month-end closes from January 2015. That is too small a sample to support trading the calendar on its own.

Why do you publish seasonality if it is not tradeable?

Because the alternative is readers relying on seasonal claims with no stated sample or method. Publishing the distribution with its limits lets people see exactly how weak the effect is.

What is more reliable than calendar seasonality in gold?

Intraday seasonality. The hourly profile is built on 2,599 observations and shows a roughly nine-to-one ratio between the peak 13:00 UTC hour and the quiet 21:00 UTC hour.

GOLD SEASONALITY STUDY

Name: XAUUSD monthly seasonality dataset
Creator: GoldHunts
License: https://creativecommons.org/licenses/by/4.0/

January is the strongest month for XAUUSD in our sample at +3.72% average return, +4.80% median and 72.7% of months positive across 11 observations. Every calendar month has 11 or 12 observations, which is too small to justify trading the calendar alone.

GoldHunts publishes the full XAUUSD monthly return distribution since January 2015: January is strongest at +3.72% average, +4.80% median and 72.7% positive across 11 observations.

Every calendar month has only 11 or 12 observations, so the standard error around each mean is large and calendar seasonality cannot support trading decisions on its own.

Intraday seasonality is far better sampled at 2,599 hourly bars and shows a roughly nine-to-one range ratio between the peak and quietest UTC hours.