Collabora Office POISSON: Exact-count and cumulative Poisson probabilities
Use legacy POISSON with its default cumulative mode, count flooring and positive-mean restriction in the pinned Linux desktop build.
Verification status: See the evidence and test scope described in this guide. How formula examples are checked
Quick answer: =POISSON(3;2) independently gives approximately 0.8571234605, the probability of at most three events when the model mean is two. =POISSON(3;2;FALSE()) gives approximately 0.1804470443, the probability of exactly three events.
Syntax, mode and units
=POISSON(Number; Mean [; C])
Number is the event count. Mean is the expected number of events over the same observation interval. C defaults to TRUE and selects the cumulative probability P(X≤Number); FALSE/0 selects the probability mass P(X=Number). The 26.04 help calls the noncumulative result a density, but for this discrete distribution it is more precisely a probability mass.
The evaluator floors Number before calculation and rejects a resulting negative count. Mean must be strictly positive in this implementation. Although a mathematical Poisson model can be extended to mean zero with all mass at zero, =POISSON(0;0) follows an error path here.
A Poisson count model commonly assumes events occurring independently at a constant average rate over the interval. The formula does not establish that a real process satisfies those assumptions, nor that its variance equals its mean. Supply a model mean rather than treating the observed count itself as proof of the rate.
Worked example: up to three events
Enter B2 = 3 and B3 = 2. Use:
=POISSON(B2;B3)
=POISSON(B2;B3;FALSE())
For mean λ = 2, the exact-count formula is exp(−λ) × λ^k/k!. At k = 3 this becomes (4/3)exp(−2) = 0.1804470443154836.
The cumulative value sums counts zero through three:
| Count | Mass as a multiple of exp(−2) |
|---|---|
| 0 | 1 |
| 1 | 2 |
| 2 | 2 |
| 3 | 4/3 |
The total is (19/3)exp(−2) = 0.857123460498547. These are independently calculated reference values, not output from an executed Collabora workbook. Omitting C and supplying FALSE answer different questions even though the first two arguments are identical.
Advanced use: interval probability and consistent exposure
For between one and three events inclusive, subtract the zero-event probability:
=POISSON(3;2)-POISSON(0;2)
The independent result is approximately 0.7217881773, from (16/3)exp(−2). The complement of the cumulative result, 1−POISSON(3;2), instead gives the probability of four or more events, approximately 0.1428765395. This is an easy place to make an off-by-one error.
If a model rate is two events per hour and the interval becomes half an hour, its corresponding mean is one event under the constant-rate assumption. Keep rate and exposure separate in the worksheet, then multiply them to obtain Mean. A count over thirty minutes should not be evaluated using an hourly mean without that conversion.
These calculations describe the chosen model. They do not forecast a guaranteed count or validate a real-world event process from a single sample.
Boundaries, numerical limits and compatibility
=POISSON(3.9;2) uses floored count 3, giving the same cumulative reference as the main example. Prefer an explicit integer count rather than depending on truncation to repair a measurement. =POISSON(-1;2) and =POISSON(0;0) follow internal IllegalArgument.
The ScPoissonDist evaluator uses different numerical paths for large means and cumulative calculations. Comments about a particular branch’s accuracy are not a universal precision promise. Small hand-checkable examples do not certify extreme-tail subtraction or enormous event counts.
The legacy name accepts two or three arguments; POISSON.DIST requires the cumulative flag. Help also says an omitted legacy TRUE flag is inserted on saving for compatibility. That is a documented serialization statement, not a save/reopen observation from this guide. No file-format round trip was performed here.
Numerical count errors and UI error localization remain distinct: internal source error identifiers are recorded without inventing application screenshots or localized messages.
Product scope, compatibility and evidence
This guide covers the new Collabora Office desktop for Linux, using the Flathub source tag coda-26.04.4.3-1 and commit 81339d176148d0d5c81ff05d82540f58f8aaa117. The official desktop release notes and Linux package manifest identify that product/build. The package manifest is a dated observation from October 18, 2026, rather than an installed-binary check.
The 26.04 function help and exact evaluator source were inspected on October 18, 2026. Expected examples were independently calculated; no Collabora application was run. Expected scalar results were derived from the documented contracts, inspected source and independent calculations; native output formatting and error localization have not been tested. Classic, Online/CODE, other operating systems and file-format round trips require their own checks.
Availability here is established from the function-specific help and the mapped package source. Related functions may have different arguments or boundary rules. The first Collabora release introducing this function is not established by the reviewed sources. No cited retirement or deprecation notice was found. Unknown introduction dates do not imply that the function is deprecated. File-format round trips require separate verification.
Related functions
- POISSON.DIST: compare the required cumulative argument
- BINOM.DIST: use a fixed-trial binomial model when that is the intended process