Repeatability vs Accuracy: The Spec That Actually Matters for Your Robot Cell

If you have spent any time reading cobot datasheets, you have noticed that repeatability is always listed and accuracy is almost never mentioned. That is not an oversight. It is a deliberate reflection of which specification actually determines how well a robot performs in production.

Most buyers assume accuracy is the important number. It sounds more rigorous. In practice, repeatability is the specification that determines whether your automation cell works reliably cycle after cycle, and accuracy in the traditional sense is nearly irrelevant for most cobot deployments. Understanding why requires understanding what each term actually measures and how robot arms are put to work in the real world.

The Target Analogy

The clearest way to understand the difference is through a simple analogy.

Imagine a marksman firing at a target. If every shot lands in the same tight cluster, regardless of where that cluster falls on the target, the marksman is highly repeatable. If every shot lands close to the bullseye but scattered across a wide area, the marksman is accurate but not repeatable. If every shot lands in the same cluster right at the bullseye, they are both accurate and repeatable.

For a robot arm in production, repeatability is the tight cluster. The arm returns to the same position, cycle after cycle, within a very small tolerance. Accuracy would mean that position is exactly where the robot was commanded to go in absolute space.

Here is the critical insight: for most production automation, the absolute position does not matter. What matters is that the arm comes back to the same place every time.

Why Repeatability Drives Production Performance

Robot arms in manufacturing are almost never programmed by entering absolute coordinates. They are taught. An operator physically moves the arm to the correct position, records it, and the robot repeats that position on every subsequent cycle.

When a robot is taught a pick position this way, any absolute accuracy error is eliminated at the teaching step. The robot learns the specific position it was shown, not a position calculated from coordinates. What determines whether the pick is reliable is whether the arm returns to that taught position consistently — which is repeatability.

A robot with ±0.02 mm repeatability returns to within 20 micrometers of the taught position on every cycle. A robot with ±0.5 mm repeatability returns to within 500 micrometers. Both may have similar absolute accuracy in global space, but their production performance differs significantly on precision tasks.

This is why every Blue Sky Robotics arm datasheet leads with repeatability. It is the number that predicts real-world performance.

When Accuracy Actually Matters

Accuracy becomes directly relevant in two specific scenarios.

Offline programming - When robot paths are generated from CAD models or simulated environments rather than taught by demonstration, the robot needs to arrive at coordinates that were calculated rather than demonstrated. Here, the difference between where the robot is commanded to go and where it actually ends up matters directly. Poor absolute accuracy means the programmed paths do not match reality and require manual touchup after loading.

Multi-robot coordination - When two or more robot arms need to work on the same part simultaneously, or hand off a part between cells, their absolute positions in shared space need to align. Repeatability within each arm is necessary but not sufficient — each arm also needs to be positioned accurately relative to the shared coordinate system.

For the vast majority of cobot deployments in manufacturing and logistics, neither of these scenarios applies. Positions are taught, not calculated. Robots work independently rather than in tightly coordinated multi-arm cells. Repeatability is the relevant specification.

Reading Blue Sky Robotics Arm Specifications

Every arm in the Blue Sky Robotics lineup is specified with repeatability as the primary precision metric.

The UFactory Lite 6 ($3,500) achieves ±0.1 mm repeatability, covering the majority of light-duty pick and place, machine tending, and basic inspection applications where sub-millimeter cycle-to-cycle consistency is required.

The Fairino FR5 ($6,999) and Fairino FR10 ($10,199) deliver ±0.02 mm repeatability, which is 20 micrometers. That level of consistency is appropriate for precision assembly, tight-tolerance pick and place, and vision-guided inspection where the arm needs to arrive at vision-calculated coordinates with minimal mechanical error.

The distinction between these specifications matters most when the application involves tight tolerances or when the vision system is calculating pick points rather than the operator teaching them. In vision-guided cells, the camera calculates a pick coordinate and the arm needs to execute it accurately. High repeatability ensures the arm arrives where the vision system sends it, not just where it was taught.

For vision-guided cells, Blue Sky Robotics' Blue Argus platform ships the camera, compute, and vision SDK as a pre-validated system. Because the hardware is tested together, the full pipeline from object detection to pick execution performs consistently rather than accumulating error across separately specified components.

Getting Started

Use our Cobot Selector to match an arm to your precision requirements. Browse our full UFactory lineup and Fairino cobots with current pricing, or book a live demo to discuss your specific tolerance requirements with our team. To learn more about computer vision software visit Blue Argus.

FAQ

Which is more important for a robot arm, repeatability or accuracy?

For most cobot deployments where positions are taught by demonstration, repeatability is more important. It determines whether the arm returns to the same position consistently cycle after cycle. Accuracy in absolute space matters primarily for offline programming and multi-robot coordination scenarios.

What does ±0.02 mm repeatability mean in practice?

It means the robot arm returns to within 20 micrometers of the taught or calculated position on repeated cycles. For context, a human hair is approximately 70 micrometers in diameter. ±0.02 mm repeatability is appropriate for precision assembly, tight-tolerance inspection, and vision-guided applications where the arm must reliably execute coordinates calculated by a camera system.

Why do cobot datasheets list repeatability but not accuracy?

Because repeatability is what determines production performance for most cobot applications. When positions are taught by demonstration rather than programmed from absolute coordinates, absolute accuracy errors are absorbed at the teaching step. Repeatability determines whether the taught position is hit consistently, which is what actually matters in production.