FLOAT Versus FIX
Before FIX appears on a screen, a receiver very often passes through an intermediate state called FLOAT. Understanding what FLOAT represents, and why it is not simply a slower or lesser version of FIX, makes both statuses much easier to reason about in the field.
FLOAT means the receiver has not yet resolved the integer ambiguities and is instead estimating them as ordinary decimal numbers. It is an honest, lower-precision position, not a failed attempt at FIX.
An Honest Estimate, Not a Broken One
In a FLOAT solution, the receiver treats the integer ambiguities as unknown real numbers to be estimated, the same way it estimates position itself, rather than as whole numbers to be solved exactly. The resulting position is typically accurate to somewhere between a few decimeters and about a meter, a real and useful improvement over an uncorrected position, but nowhere near FIX's centimeter-level precision.
Why a Receiver Starts in FLOAT at All
Resolving integer ambiguities correctly requires enough satellites, strong enough geometry, and a short enough baseline that the receiver can be statistically confident it has found the one correct set of whole numbers among many candidates that fit the data almost as well. Until that confidence threshold is crossed, reporting FLOAT is the honest choice: a real position, clearly labeled as less precise, rather than a guess dressed up as certainty.
A FLOAT solution converging into FIX after a few seconds is completely ordinary. A FLOAT solution refusing to converge for a long stretch is itself useful information, often pointing toward multipath in a cluttered environment, poor satellite geometry, obstructed sky view, or a baseline stretched further than the correction service can support well.
FLOAT is not simply FIX that has not finished loading. It is a distinct type of solution with its own accuracy characteristics, and a FLOAT position should never be treated as centimeter-level just because FIX is expected to arrive shortly.
Why This Distinction Matters for Trust
A crew that has internalized the difference between FLOAT and FIX asks a very different question when a point looks slightly off: was this actually collected under FIX, or did the collector record a FLOAT position during a brief dropout, without anyone noticing the status at the moment of collection. That single confusion, a FLOAT point silently recorded as though it were FIX, is a common and entirely avoidable source of field error.
A point in a dataset sits noticeably off from its neighbors, all collected during the same site visit. Before suspecting the base or the reference frame, what is worth checking first? Whether that specific point was actually collected under FIX, or whether it was recorded during a brief FLOAT interval that went unnoticed at the time.
FLOAT is an honest, decimeter-to-meter-level position produced when integer ambiguities have not yet been resolved. It is a different kind of solution from FIX, not an inferior version of it, and confirming which status a point was actually collected under is a basic, essential verification step.
Even a genuine FIX solution is not automatically the accuracy a project actually needs. The next chapter draws a distinction that sits at the center of this entire companion.