How we rate robotic pool cleaners

What the score is built from

Published maker specs (company claims), prices we saw on the maker's US store or a major US retailer, and owner reports we summarized from public threads and published reviews. We do not test robots in our own pool. That last sentence matters. A lab score tells you how one robot did in one pool on one day. Ours tells you what a robot is built and sold to do, on the same scale for every model, so you can compare a $300 floor cleaner with a $2,300 skimming robot without either of us guessing.

Each criterion scores 0 to 10 from fields in data/pool_robots.json. The overall RobotDoorstep score is the weighted average times 10, so it runs 0 to 100. Maker numbers such as runtime, pool size and filter fineness are company claims, and we label them that way everywhere they appear. Prices come only from the maker's own US store or a major US retailer, with the date we saw them. Owner reports are our summary of recurring themes in public owner threads and published reviews, linked from each robot's page; they are not a survey.

What doesn't touch the score: store links, commissions, review samples and advertising. A robot with no store link at all can top the table.

The eight criteria and their weights

Coverage and pool fit carry the most weight on purpose. Two complaints come up again and again in the owner threads we summarized, and neither is about a weak motor: a ring of grime the robot never reaches at the waterline, and a battery that quits before a pool it was never rated for is done. Price only enters through the value criterion, so an expensive robot can still score well if it does more.

Exactly how points are given

Each criterion is scored 0 to 10 from fixed bands. Nothing is adjusted by hand: change a field in the database and the score changes on the next build.

Cleaning coverage (20%)

Pool fit (15%)

Battery and runtime (10%)

Filtration (15%)

Ease of maintenance (5%)

Value per dollar (15%)

Warranty and support (10%)

Owner-reported reliability (10%)

A worked example

Here is the full breakdown for the Beatbot AquaSense 2, the highest-scoring robot in the database today. Every line comes straight from its row in the comparison database.

Notice what keeps it from a perfect score: no surface skimming, a standard 150 micron basket rather than an ultra-fine one, and a weight above 20 lb. Robots that beat it on one of those give up points elsewhere: the Beatbot Sora 70 adds surface skimming but costs $300 more, and the Seauto SAT40 lists a 10 micron filter but Seauto doesn't state a warranty length on the product page we read.

Rules for picks, ties and missing data

• Minimum data. We only score a robot when the maker publishes at least three of four core specs: what it cleans (floor, wall, waterline, surface), maximum pool size, warranty length and runtime (or that it is corded). Robots below that line are listed in the database as not scored, because guessing the missing specs would be our opinion dressed up as data.
• No price, no value score. A robot with no price we could see on the scan date gets a provisional score: the value criterion is left out and the other weights are scaled up. Provisional robots can't win a best-for pick.
• Best value. Best value only considers robots scoring 70 or more overall, so a cheap robot with weak specs can't win on price alone.
• Who can win a pick. Best-for picks only consider robots with status current, a price we saw on the scan date, and no sold-out, unavailable or pre-order flag. Surface skimmers are listed but not scored against floor robots.
• Best-for picks. Picks are computed in a fixed order and each robot can win only one: best overall, best for above-ground pools (maker lists above-ground use and a published weight of 20 lb or less), best under $500, best value, best for large pools (3,000+ sq ft), best for waterline and tile, best for leaves (surface skimming), best for screened Florida pools (finest published filter among waterline cleaners). The current winners are in our buying guide.
• Ties. Equal overall scores: lower price first, then longer warranty.

Robots we don't score yet (32)

These robots are in the database with every figure their makers publish, but not enough of it to score fairly. Several are well-known corded models whose makers sell through dealers and publish little on their own sites. If a maker publishes the missing spec, the robot gets scored on the next update.

Another 12 robots have a score but no price we could see on a maker store or major retailer, so their value criterion is left out and they can't win a pick: Maytronics Dolphin Liberty 400, Maytronics Dolphin Liberty 600, Maytronics Dolphin Oasis Z5i, Polaris 9450 Sport, Polaris ERA 710, Hayward The PoolCleaner R110 (Open Line), Hayward The PoolCleaner R130 (Open Line), Hayward SharkVac, Hayward SharkVac XL (Open Line), Hayward TigerShark, Hayward TigerShark QC, Hayward AquaVac 250Li.

How the score is kept current

The score is versioned by edition. The 2026 weights and bands above apply to every robot until the next annual review, so a score from March and a score from October mean the same thing. Between editions we add new models as soon as a maker publishes US specs, retire discontinued ones to a "discontinued" status (they stay in the database for anyone shopping used), and refresh prices every month in the pool robot price tracker. Scores recompute on every build, and the build refuses to publish if a stored score doesn't match the formula.

At each annual review we check whether the weights still match what owners complain about and what new robots can do. If a band changes, the new edition says what changed and why.