Bitcoin Outputs per Transaction (P90)¶
What It Measures¶
Bitcoin Outputs per Transaction (P90) shows the 90th-percentile output count of Bitcoin transactions on a given day.
It answers the upper-tail version of the output-side question:
How output-heavy were transactions near the top end of the daily distribution?
A 90th-percentile value means that 90% of transactions created this number of outputs or fewer, while the upper 10% created more.
This is the tail-sensitive output-count metric in the family.
How To Use It¶
This metric is useful when you want to inspect upper-end transaction complexity on the output side.
It helps answer questions such as:
- Is the high-output tail expanding?
- Are more complex multi-output transactions becoming more important?
- Did the average rise because the upper tail stretched, or because the whole distribution shifted?
This series should be read together with:
- Average Outputs per Transaction
- Median Outputs per Transaction
Within the family:
- Average is the canonical baseline,
- Median shows the typical transaction,
- P90 shows what is happening near the complex-output tail.
What It Can Say About Market Regime¶
P90 output count is not a market-cycle signal. It is a transaction-structure signal.
Rising P90 with stable median¶
This usually means the upper tail is stretching while the typical transaction is broadly unchanged. More complex output structures are becoming more visible, but mostly at the high end.
Rising P90 with rising average and median¶
This suggests a broader structural shift in transaction composition. Output complexity is no longer confined to the tail.
Why P90 matters here¶
Without a tail metric, it can be hard to know whether rising average outputs reflect genuine broad change or just a minority of more complex transactions. P90 gives the cleanest direct read on that upper-end behavior.
Historical Background¶
Percentile-based output metrics became useful for the same reason percentile-based input metrics did: Bitcoin transaction structure is heterogeneous, and averages alone do not tell you whether the change came from the center or from the tail of the distribution.