Ask which filament is best for an outdoor bracket and you will get PETG, ASA and PETG-CF from three people who are all correct.
They are correct because they are answering three different questions. Nobody stated the machine, nobody stated how much difficulty was acceptable, and those two things decide the answer before any material property gets a vote.
State the requirement as weights, not adjectives
"Strong" is not a requirement. Strong in tension or strong in impact? At room temperature or in a car in July? Once or ten thousand times?
The material selector forces the issue by making you weight the criteria against each other. For the bracket above the brief looks like this: outdoor exposure weighted 5, mechanical strength 3, heat resistance 2, ease of printing 3, cost 2. Nothing exotic — a part that lives on a fence, takes some load, never gets hot, and that you would like to print without a fight.
That is already more specification than most material choices ever receive, and it is what makes the comparison possible at all.
The machine eliminates before the ranking starts
Run that brief for an open-frame printer with a brass nozzle and a 260 °C hotend, and the selector considers 20 of the 35 materials on the site. Fifteen are gone before any scoring happens: everything that needs an enclosure, everything that needs a hotter nozzle, everything abrasive enough to eat the nozzle.
The ranking that comes back is PETG at 0.6941, then PETG-CF at 0.6304, then PET at 0.6196.
Now change one thing about the workshop and nothing about the part. Enclosed machine, hardened nozzle, 300 °C hotend. The eligible set jumps to 33 of 35, and ASA — the only material on this site rated excellent rather than merely good for UV, which is exactly what an outdoor bracket wants — appears at 0.6696.
It was never a worse choice on the open-frame machine. It was an impossible one. That distinction is missing from almost every "best filament for X" list, and it is the difference between advice and a shopping recommendation.
Ease of printing is a requirement, not a personality trait
Notice something about that second run: ASA arrived in the list but PETG still won, on an identical score to before.
That is because ease was weighted 3, and ASA is genuinely hard — it warps as readily as ABS, and the enclosure that made it eligible does not make it pleasant. Drop the ease weight to 1, meaning "I will accept a difficult print for the right material", and the answer flips: ASA takes first place at 0.7342, on the strength of the one criterion the part actually cares about.
Same part. Same machine. Same five criteria. The only thing that changed was how much trouble the operator was willing to accept, and that is what decided it.
This is why forum answers diverge so violently and why both sides are convinced the other is wrong. Half the room is optimising for the part and half is optimising for the evening, and neither says so. If you are new, weight ease high on purpose and know that you are doing it — the easiest filaments to print exist, they are a legitimate choice, and choosing one is not a compromise if the part does not need more.
Read the "bad at" list, not the "good at" list
Every material page here stores an explicit list of what that filament is genuinely bad at, and the selector prints it alongside the winner rather than underneath a fold. For the bracket, the top pick's list reads: droops on overhangs, strings unless retraction is tuned, does not glue well.
That last one has ended more projects than any tensile figure. It is not a weakness anyone markets, and it does not appear in a strength ranking, but if your bracket was going to be bonded to something it is disqualifying.
A useful habit: pick your material from the ranking, then read only the bad-at list, and ask whether any line on it describes your part. If one does, take the next material down. That single step catches most of the choices people regret, because reputation is built from what a material does well and disappointment comes from what it does badly.
Requirements that are not requirements
"Strong." The strongest filament on tensile numbers is a fibre-filled nylon that needs a 300 °C hotend, a heated chamber, a hardened nozzle and a dry box, and that is brittle under impact compared with cheaper alternatives. If you cannot state which kind of strong, the answer will be wrong in a way that looks right on a datasheet.
"Heat resistant." Heat resistance is a continuum, and the honest question is what temperature the part will actually see. A part in a car interior and a part next to a hot-end mount are two entirely different specifications, and one of them is satisfied by a material you already own.
"Food safe." This is the one where the material is the smallest part of the problem. Layer lines harbour bacteria regardless of polymer, and the nozzle, the moisture and the additives are all in the food path too. Whether a filament is food safe is a question about the whole process, and the honest answer for most printed parts is that the polymer's certification does not transfer to your part.
"Cheap." Spool price is a weak proxy for cost. A cheap material with a high failure rate and a long print time is not cheap, and cost per finished part is a different number from cost per kilogram.
What to do with the answer
The selector returns a ranking, a per-criterion breakdown showing where each score came from, and a stated tradeoff naming the winner's weakest criterion. Use all three. A score is only interesting if you can see which weighting produced it, which is why the contributions are printed rather than summarised.
Then check two practical things the ranking cannot know. Whether you can actually buy the material locally in the colour the part needs, and whether it needs drying you are not set up to do — several of the nylons and every fibre-filled grade expect to be fed from inside a heated enclosure, not simply dried beforehand, and PA6-CF will disappoint anyone who runs it once and then leaves the reel on a shelf.
For the outdoor case specifically, PETG against ASA is the comparison worth reading in full, and which filament is best outdoors is the short answer with its assumptions stated. If your part is a gear rather than a bracket, the gear question weights things quite differently, and an enclosure changes what is available to you far more than any spool purchase will.