I tapped three notes into a reverse chord lookup tool.

C on the low E at fret 8. Open A. G on the D string at fret 5.

Three notes. A, C, G.

It came back with five names. The one it put first - biggest text, top of the list, with a small red warning triangle beside it - was C6(no3).

A C major sixth chord. With no third. Ranked above Am7. On a voicing whose lowest note is an open A.

None of the five are strictly wrong. Every one contains A, C and G. But the honest answer to "what am I playing?" - with an A in the bass - is Am7. The tool had Am7 in the list. It just didn't lead with it.

TL;DR Reverse chord lookup isn't a lookup problem - it's a ranking problem. Hand three notes (A, C, G) to a typical reverse chord lookup and it leads with C6(no3) - a third-less sixth chord - over the obvious Am7, then flags its own top answer with a warning. Fretscape is root-aware: it defaults the root to the bass note, leads with Am7, and keeps the exotic readings one tap away, each labelled with exactly what's omitted or inverted. Same notes, same candidate names - better order, honest labels.

The failure isn't the list. It's the order.

Reverse lookup tools fail in two obvious directions, and most pick one.

The too-much failure: solving a set membership problem. "Which chord definitions include these notes as a subset?" For any three notes, dozens qualify; for four, some tools will hand you forty labels. A guitarist looking at those notes isn't hearing forty chords. Listing every superset isn't help - it's noise dressed as knowledge.

The too-little failure: hand the same voicing to a different tool and get "no chord found" - not because it's unnameable, but because the tool never learned how to name it.

This is neither failure - and that's the point. This isn't a careless tool; reverse lookup is genuinely hard, and ranking C6(no3) first is the call plenty of tools make. Five names is a reasonable list. Am7 is right there in it. The failure is subtler, and more interesting: the right name is in the list, and it isn't on top. A valid label, wrong headline. The warning triangle is the tool's own admission that its first answer has no third - and it ranks it first anyway.

A chord name isn't a lookup. It's a claim about what you're most likely holding. The whole job is deciding which claim goes first.

What a chord name actually means

A chord name is an identification. "This sound, that name."

Identification needs a reference point, and on guitar that reference point is usually the bass. Play A, C and G with an A underneath everything and you are, to almost any ear, playing some flavour of A minor. Calling it C6(no3) asks the listener to hear a third-less C chord inverted so its sixth (the A) sits in the bass - a reading that's available, but not what's happening.

Some tools treat your input as a query against a table: "show me everything these notes belong to, likeliest guess first." The hard part was never finding the matches. It's ordering them like a musician would.

Root-aware, by default

Here's the same three notes in Fretscape.

The root selector sits on Auto, and auto means the lowest-pitched note - here, the open A. So Fretscape leads with Am7, shows you which note is doing what - root (A), third (C), seventh (G), fifth omitted - and offers C6/A right below it as an alternative, tagged plainly: inversion, no 3rd. The exotic reading isn't hidden. It's just not the headline.

Pivot the root yourself and Fretscape follows you honestly. Set root to C and it leads with C6/A - C6(omit3)/A - and drops Am7, because Am7 isn't a C-rooted chord. Set root to G and it surfaces Gsus2sus4/A and Gsus4(add9)/A - purely because those are the only valid G-rooted readings, and a tool that's been asked a question shouldn't refuse to answer.

The difference isn't the candidate pool. It's that Fretscape decides what to put first the way a guitarist would - from the bass up - and labels every alternative with what it actually is.

The quiet one

Here's the kind of detail that only shows up when you stress-test naming across hundreds of voicings.

Both tools can name the G-rooted reading of A, C, G. The other one calls it Gsus(add9) - but open the detail and it's really Gsus(add9)(omit5): start from a sus4, bolt on a 9th, then drop the 5th to make the notes fit. Three contortions to name three notes.

Fretscape names the same notes directly: Gsus2sus4(omit5) - a 2 and a 4 both present, no third, fifth omitted. And it has the Gsus4(add9) reading too; it just ranks it second, behind the cleaner spelling. The same decision as C6(no3)-versus-Am7, one level down: from the same valid options, pick the one a player would actually reach for.

By default you never see any of this. Those obscure G-rooted spellings live behind Theory Mode - off unless you ask - because they're not useful to most people most of the time. Flip Theory Mode on and they appear in the normal view too, clearly marked as the exotica they are.

Complexity isn't a property of the name

This is where the design got genuinely hard.

E7#9 looks complex written down. Altered tension, implied seventh, root note - a lot going on for four characters.

But for the Hendrix voicing - the iconic E / G# / D / G shape - E7#9 is the simplest accurate label. The voicing literally is that chord. Anything shorter would be missing information.

Now picture that same spelling on a voicing where "E major" would fit just as cleanly. Suddenly E7#9 isn't clarity. It's noise.

So "complex spelling" isn't a property of the name. It's a property of the pairing - this name, against this voicing. The same label is the clearest possible answer on one shape and pure showing-off on another.

E7#9 is complex in a vacuum. On the Hendrix voicing, it's the clearest thing you could possibly call it.

And sometimes a "valid" name is one no voicing should ever earn.

Take x7888x - four notes, E, Bb, Eb, G. One valid label is Em(maj7)b5: minor, major 7th, the 5th moved down a semitone. Another is Emmaj7(#11): minor, major 7th, a #11 sitting above a fifth that isn't in the voicing. The first matches the actual notes. The second invents a fifth to justify itself.

Fretscape shows you Em(maj7)b5. It doesn't show Emmaj7(#11) - not by default, not behind Theory Mode, not anywhere. A label that invents a note to exist doesn't deserve a place in the list. A common approach puts it at the top.

Partition, not penalty

When an exotic reading is legitimate - a real jazz spelling, a genuine rootless voicing - Fretscape doesn't bury it on a sliding scale beneath the plain names. It puts it in its own room, behind its own switch, clearly labelled.

Theory Mode is one of those switches. "Show rootless voicings" is another. Off by default, both of them. When a switch is off, those labels aren't demoted - they're simply not in the set. When it's on, they don't compete with the standard names point-for-point; they sit in their own bucket, and you always know which kind of label you're looking at.

Partition, not penalty. The easy labels don't fight the exotic ones on a sliding scale. They live in different rooms, and you decide which rooms are open.

The rootless problem specifically

Jazz guitarists leave roots out all the time - a bassist or pianist supplies the root. A rootless Am7 can absolutely still sound like Am7 if the rest of the chord's identity is there. So "no root" isn't an automatic disqualification.

But it raises the bar. If you're going to call something a rootless something, the chord's identity has to be present. A rootless minor-seventh has to carry the notes that make it recognisably minor-seventh. Missing those, the voicing hasn't earned the label - it's just a fragment wearing a confident name.

Leaving the root out is fine. Leaving the root out and the chord's identity is just a few notes and a hopeful label.

And some families aren't candidates for rootless naming at all. Triads collapse into simpler fragments without the root. Suspended chords are already missing the third - pull the root and there's barely a shape left. Power chords are literally "root + 5th." It isn't a technical limitation; it's a refusal to promise a label you can't defend.

The user never sees the sausage

Default behaviour: the honest, root-aware answer first, plain alternatives beneath it, and the exotica tucked behind Theory Mode and the rootless switch. Two opt-ins, independent, for the players who actually want them.

Even with everything switched on, only chords that earn their label appear. Nothing is flagged just to make the list look bigger. The chord you're most likely holding comes first. No warning triangle on your own headline answer. No wall of names. No guessing game.

You can try it yourself in Fretscape's chord identifier - tap in the notes you're holding and see the names a voicing actually earns, in the order a guitarist would expect them.

Why this is harder than it sounds

"Show every possible name" is a five-line algorithm. "Show the names a player would actually hear, in the order they'd hear them" is a design document.

Whether a spelling counts as exotic depends entirely on what simpler label is available for this specific voicing - the same spelling that's noise on one chord is the only honest name for another. The rules can't sit on the name. They have to live with the voicing.

Reverse lookup quality is invisible when it's good, and only the noise shows when it's bad.

Quantity is cheap. The right answer on top is a decision you have to keep making.

The broader pattern

This is the fourth of five How Fretscape Thinks posts, and Fretscape has made the same call every time. The Am → F transition story - the system understood physical movement better than my intuition did. The progression optimiser bugs - the system wasn't respecting user intent. The voicing ranking fixes - the system wasn't respecting user expectation. And now this - a reverse chord lookup had the right answer in hand and led with the wrong one. (One more after this: a key detector that knows when it doesn't know.)

Every time, the same question: what does a guitarist actually want here? Every time, the same answer: not the impressive-looking maximum. The honest one, first.

Breadth is easy. Depth shows up in the order.

A tool that leads with C6(no3) and warns about it is performing knowledge. A tool that leads with Am7 and keeps the rest one honest tap away is giving you knowledge.

I'd rather be useful than impressive.

A few questions this raises

Isn't C6(no3) a valid name for those notes?

Yes - that's the point. It's valid, and it's still the wrong thing to lead with. Naming a chord isn't about finding labels that fit; dozens fit. It's about ranking them the way a musician would, from the bass up. A C major sixth with no third, sitting on an open A, is a reading you can construct - not the one you're hearing.

Why does Fretscape default the root to the bass note?

Because on guitar the lowest note does most of the work of telling your ear what the chord is. Auto-rooting to the bass is the closest a tool can get to hearing the voicing the way you do. You can override it in one tap - and when you do, Fretscape re-ranks honestly around your choice instead of pretending nothing changed.

What about players who actually want the exotic spellings?

That's what Theory Mode and the rootless switch are for. Turn them on and the jazz and theory-heavy readings appear, in their own clearly-marked partition - they don't fight the standard labels on a sliding scale. Anyone who knows what they're looking for can have it. Beginners who don't aren't subjected to it.

What if a voicing has no clean name at all?

Then Fretscape says so. A reverse lookup confident enough to come back empty is more useful than one that confabulates a label to fill the space. "Not enough information to name this" is honest. "It's a Bm7b5/A if you squint" is performance.