The basics take about two minutes to learn. Getting good at it takes longer. This guide covers both: the mechanics, then the actual strategy that separates players who stumble through on guess four or five from players who regularly get it in one or two.
On this page:
Each day Kpopless gives you three song puzzles. For each puzzle, you listen to a short audio clip from a K-pop song and try to guess the title before you run out of attempts. The clip starts at 1 second and gets longer with each wrong guess or skip. You have six attempts per song.
The puzzles reset at midnight London time (GMT in winter / BST in summer). That's 7 PM Eastern, 4 PM Pacific, 8 AM Korean Standard Time. Everyone worldwide gets the same three songs on the same day.
Here's exactly how long each clip is, and honestly what you can realistically identify from each one:
Honestly, not much, unless you know the song well. At one second you're mostly getting a texture: the synth type, whether there are vocals or an instrumental intro, the general tempo. Players who get songs on the first guess have usually recognised that opening fragment from pure memory. You either know it or you don't. Don't agonise over guess one if you have no instinct; save the commit for later.
Two seconds starts to give you structure: verse, chorus, something in between. At four seconds you can often catch a full vocal phrase or a hook, which is where era recognition actually kicks in. You're no longer just hearing texture, you're hearing something you can work with.
By 7 seconds you should know the group, even if you haven't landed the specific song yet. Ten seconds, which is what both the 5th and 6th guess give you, is usually enough for the hook. If you genuinely still can't identify something at 10 seconds, it's almost certainly a B-side. Accept that, take your best guess based on what you know about the likely artist, and move on.
Skipping is free. It costs no guesses and advances you to the next clip length. This is one of the most misunderstood mechanics for new players; many people treat a skip like a failed guess, but it isn't.
No instinct at 1 second? Skip, don't guess. You haven't lost anything. Same if you've narrowed it to an artist but can't figure out which song: another clip might give you the hook you need. The rule is: if you're not actually confident, there's no reason to spend a guess to find out you were wrong.
If you're reasonably sure (60%+), go ahead and guess rather than skip. A wrong guess and a skip both advance you to the same next clip, but the guess costs an attempt. So the break-even is confidence. Also: if you've locked down the artist and you're choosing between two songs, a 50/50 on guess 3 or earlier is worth taking.
Some players skip every time until they reach the 7 or 10-second clip, then guess. This is a valid strategy but it caps your score: you'll never get a 1 or 2 on those songs. If you're playing for fun that's fine. If you're trying to post a competitive result, you need to commit earlier.
K-pop has a distinct sonic fingerprint for each generation. Learning to place a song's era quickly narrows your search from 2,000+ tracks to a few hundred. This is the single skill that makes the biggest difference between a player who plateaus at guess 4 and one who regularly gets songs in 2.
H.O.T, g.o.d, S.E.S, Sechs Kies, Fin.K.L. Honestly the easiest era to clock once you've heard enough of it: there's something about the production that just sounds like a different decade, and it's not subtle. Tinnier synths, simpler drum patterns, the way vocals sit against the instrumental track. Pre-autotune idol vocal style is its own thing that doesn't really show up later. When you're not sure if you're dealing with 1st gen or just a cheap-sounding 2nd gen track, guess 1st gen. You'll be right more often than not.
If you were into K-pop around 2009-2013, this era lives in your memory. BIGBANG, Girls' Generation, SHINee, 2NE1, f(x), groups with identities so specific that a few seconds is usually enough. GG sounds like GG. SHINee sounds like SHINee. That distinctness is genuinely useful for the game. Production got bigger and more confident, EDM started creeping in around 2011. Anything that feels polished but clearly predates the tropical house wave is probably sitting somewhere in 2nd gen territory.
Genuinely the hardest era to learn because there's no single 3rd gen sound. BTS and TWICE and EXO and Red Velvet sound nothing like each other. What they do share is money: the budgets got serious, the production got global-facing, and K-pop started showing up on Western radio and streaming charts around 2017-2019. Dense, expensive-sounding. If it feels like the kind of K-pop that broke into playlists you weren't expecting to find it on, you're probably here. The 2019-2021 window is where 3rd and 4th gen blur, and that transition period will catch you out regularly until you've heard a lot of both.
aespa, IVE, LE SSERAFIM, NewJeans, ENHYPEN, TXT, (G)I-DLE. The low end hits harder, song structures get stranger, and there's a lot of deliberate era-hopping in the references: NewJeans went full Y2K, aespa went synth-pop, LE SSERAFIM just goes hard and doesn't really explain itself. IVE keeps it clean commercial pop. Stray Kids and TXT sit somewhere slightly apart from all of this, darker and more rock-influenced, not really the same wave as the girl-group-led mainstream 4th gen sound. If it sounds like something built for 2023 TikTok, you're here.
No shortcut here: your era recognition does the work. Use the first one or two clips purely to place the era and gender of the vocalist(s). Mixed-gender groups almost don't exist in mainstream K-pop, so if you can tell "male vocals" or "female vocals," you've already halved the search space. Then narrow by era, then by group sound, then by song.
The pool is still large but more manageable. The big differentiators are production style and vocal character. BTS solo-era and late BTS have a very different sound from mid-era BTS. EXO has a distinctive vocal layering that's hard to mistake. Stray Kids have a specific self-produced rawness. ATEEZ lean into theatrical, high-drama production. SHINee's older catalogue has a funk and R&B flavour most 4th gen groups don't replicate. If you hear something understated and emotionally restrained, check VIXX or B.A.P. If it's maximalist and anthemic, check ATEEZ.
This mode rewards knowing your groups' sonic identities cold. BLACKPINK have arguably the most distinctive sound in K-pop: the contrast between Jennie and Lisa's rap sections and the chorus production is instantly recognisable. NewJeans don't sound like anyone else in the current generation. IVE and LE SSERAFIM are the hardest to distinguish from each other for many players since they occupy similar sonic territory, though LE SSERAFIM tends to lean harder and IVE tends toward cleaner pop. aespa are distinctive once you've heard a few of their tracks, with dense synth textures and clear vocal separation. TWICE's older catalogue is bubblegum pop; their recent output is darker and more mature. MAMAMOO are always a safe guess if you hear confident, powerful, slightly jazzy vocal performance.
This one is fundamentally different. See the Album Guess section below for dedicated tips.
The autocomplete searches every song and artist name in the library. A few things that aren't obvious:
Spelling doesn't need to be exact; the search handles romanisation variations and minor typos, so "dynamit" finds Dynamite, "gee" finds Girls' Generation, that kind of thing. If you know a song by its Korean title, that works too. Both versions are in the database for bilingual songs.
One of the most useful techniques: if you know the artist but not the specific song, just type the artist name and browse through their tracks in the dropdown. You'll see everything they have in the library, which makes a 50/50 guess much more informed.
Watch out for shared song titles: a few common words appear across multiple artists, so the dropdown might show several results. The artist name is always shown alongside the title, so just confirm you're picking the right one before submitting. Album Guess uses the same search but for album names; same logic applies.
Easily the most common one. You hear something vaguely familiar, panic a bit, guess, get it wrong, and now you've spent an attempt and the clip length is still only at 1 second. If you're not reasonably confident, just skip. It's free. There's no award for committing early on a shaky instinct.
The library has 2,000+ songs. If you start guessing specific titles without narrowing by era and gender first, you're working with the whole pool at random. The better approach is always broad to specific: era first, then gender, then group, then specific song. That sequence makes the search manageable.
Some players treat every skip as a moral failure. It isn't; skipping is part of the game. A player who skips once and gets the song right on guess 2 has done better than a player who guesses wrong twice and gets it on guess 3. Same information consumed, better outcome.
Late 3rd gen and early 4th gen overlap in ways that will trip you up until you've heard a lot of both. This isn't really a fixable mistake so much as something that improves with time. The K-Pop Generations guide is worth reading if you keep getting this window wrong; it goes into the specific sonic differences in detail.
EXO's D.O., SHINee's Onew, BTOB's Sungjae. High-tenor male K-pop vocalists exist and they fool people regularly. If you're in General mode and genuinely can't tell the gender from the vocals alone, wait for a longer clip and look for production cues instead; the instrumentation and arrangement often give more away than the voice does.
Album Guess works on the same six-guess structure but with album cover art instead of audio. The cover is shown in six sections, each guess revealing another piece. You're trying to identify the album (artist and album name) before the full image is visible.
Colour palette first. K-pop albums have strong visual identities: HYBE artists tend toward specific palettes within a discography era, YG albums are usually darker and more graphic-design-heavy, SM albums have their own recurring aesthetic language. You can often narrow down the label before you can see any faces or text.
Era matters here exactly like it does for audio. A cover using the visual language of 2015 K-pop (specific filter styles, typography choices, styling) narrows your search significantly before you can even see the whole thing. Styling is era-locked: what BTS's hair looked like in 2018 is different from 2021, and K-pop fans typically remember which era a look belongs to.
Don't wait for the full reveal. If you need all six fragments to make a confident guess, you've already burned most of your attempts. Try to commit on three or four pieces. The full image being visible is a last resort, not the goal.
One slightly annoying thing about Album Guess: some iconic albums that you'd recognise in a millisecond as a full cover are genuinely cryptic in fragments, while some obscure ones are oddly guessable from a specific corner. The difficulty is partly about which fragment reveals first, which isn't something you can control.
Your score for each song is the number of the guess you got it on. A 1 means first-guess, a 2 means second-guess, and so on. X means you didn't get it within six attempts. Lower is better.
Your three-song result (for example, 2/1/4) is shown on a result card alongside your guess distribution compared to all other players that day. The distribution shows you, for each song, what percentage of players guessed on attempt 1, 2, 3, and so on. A song where 80% of players got it in one guess was clearly recognisable; a song where most players needed 4-5 attempts was a genuinely hard one.
The shareable version of your card uses blocks rather than numbers, so you can post it without revealing the song titles, the same approach Wordle popularised. Players on Twitter and Discord post these regularly; the game is designed so that discussing your scores with other players the same day is part of the fun.
Results are saved in your browser and persist until the next reset at midnight London time. They're not sent to any server; there's no account system. Clearing your browser data will erase your saved results.
Head back to the game and put some of this into practice. The best way to get better is just to play consistently; your era recognition improves quickly once you start noticing what you got wrong and why.