Signals

Sixteen Novembers

Published 21 November 2026

Accuracy compared with sample size and window

In The Economic Times’ update at 12:17 p.m. IST on 21 November, with Bitcoin quoted at $56,233, Prateek Gupta of Mudrex said Bitcoin had finished higher in 10 of the previous 16 Novembers. The same comment said this November was down over 20% and was a seventh straight weekly loss. Below the summer low already set that seasonal line beside the losing streak. The question left open is what the 16 contains.

One row per November

A calendar count gives every November one observation, whatever the assistant would have called. The denominator is 16 months. The numerator is how many of those months ended above where they began, on the definition Mudrex used. The count does not say which coins, which timeframe, or which published call was being scored. It does not say whether a Hold, a Buy, or a Sell was on the card in any of those months.

The current month is outside the 16. Gupta called the previous 16 Novembers, and then described this November separately, at the $56,233 print in that update. Adding this month to the 10 would mix a month that has not closed into a count of finished months. Leaving it out is what the sentence already does.

What an accuracy report counts instead

How to read crypto signal accuracy reports starts from a different denominator. Directional accuracy is how often a prediction matched the later price direction. The report has to say how many predictions were included, the date range, which coins and timeframes were in scope, and whether fees or exclusions applied. Ten trades can look perfect and still mean nothing, because ten is the sample, not a season.

A month that finished higher is not a prediction that was right. A month that finished lower is not a Hold that failed. Until the row names the call, the time it was published, and the price it was scored against, it cannot sit in the accuracy numerator. Counting calendar months tells you how often November, as a month, closed up. It tells you nothing about how often a particular card call was right.

Beside a CryptoCrispy plan

Use the seasonal sentence as a description of other Novembers, attached to the print it was said with. The call on the card stays the call that was published for this window.

Keep the assistant call that is already published. Keep the position size printed for that trade. Hold adds nothing. Count the call once, even when the Pivot agrees.

Takeaways

  • At the $56,233 print on 21 November, Mudrex’s count was 10 higher Novembers out of the previous 16.
  • Each of those 16 is one month, with no assistant call attached.
  • An accuracy report counts predictions, the window, and the coin set. A calendar count does not.

CryptoCrispy provides market analytics and trading tools. This article is general information, not personal financial advice. The November count is Mudrex’s, via The Economic Times on 21 November 2026. It is not CryptoCrispy performance.