Debutle

The title track is not a free pick

By the person who builds Debutle · Last updated 24 September 2026

Someone on Hacker News asked me to drop every album that has a song with the same name. In Expert, where ratings are hidden, a track named after the album looks like the answer printed on the card. So I counted. It is the best song on its album less than one time in five.

The worry is fair on its face. Expert shows you song titles and nothing else: no rating, no single tag. Spin AC/DC's Back in Black, see a song called Back in Black, and you can take it without knowing a thing about AC/DC. It is rated 96, the best song on the record. If that held across the pool, Expert would have a hole in it.

How many albums have one

137 of the 463 albums in the pool have a song that shares the album's name. That is 30 percent, so a ten-album Daily usually deals about three. I left out six more where the only match is an intro, like The Chronic (intro) or Life After Death (intro). Those sit at 62 under the short-form cap, and nobody mistakes them for the hit.

How often it is the right pick

24 times out of 137, counting ties. That is 18 percent. The typical title track is the fourth-best song on its album, and on average it sits 9.4 points below the album's best. A song picked at random from the same albums sits 12.0 points below. So the name is a hint, but a weak one: it buys you under three points over closing your eyes.

title track is the best song on its albumall 137 title tracks24 of 137 · 18%counted as singles (56)21 of 56 · 38%not singles (81)3 of 81 · 4%
The hint mostly works when the title track was already a single you would have known anyway.

Some very famous title tracks lose. Thriller is 94 and third on its album, behind Billie Jean at 98 and Beat It at 96. Bad is 91 and fourth, behind Smooth Criminal. Born in the U.S.A. is 89 and third, behind Dancing in the Dark and I'm on Fire. Katy Perry's Teenage Dream is 92, a great number, and still only fifth on an album where four songs sit at 94 or 95.

Where the hint works, and where it is a trap

Split the title tracks by whether the game counts them as singles and the picture changes. The 56 singles are the album's best song 21 times, 38 percent, and on average sit only 4 points below the top. That is where the Hacker News worry is real: Back in Black, ...Baby One More Time, Starboy, Toxicity.

The other 81 are the album's best song three times. On average they sit 13.1 points below the top, which is worse than picking blind. Plenty of albums are named after a quiet track or an instrumental. Pet Sounds is 68, twelfth of thirteen, on the album that has God Only Knows at 96. ABBA's Arrival is 67 and dead last, on the album with Dancing Queen. The Miseducation of Lauryn Hill is 73, thirteenth of fifteen. Drake's Views is 79, nineteenth of twenty, on the album with One Dance. Two of the three exceptions are Pink Floyd's Wish You Were Here at 93 and 2Pac's All Eyez on Me, tied with Ambitionz az a Ridah at 92.

Genre and era

Pop is where title tracks are common and often right: 68 of 159 Pop albums have one, and 15 of those are the album's best song. Rock hits at a similar rate, 8 of 39. Hip-Hop barely plays: 30 of 158 albums have a title track, and only All Eyez on Me tops its album. If you draft Hip-Hop in Expert, the hint is close to useless.

By era, the hint is best on old records: 10 of 26 title tracks from before 1990 are their album's best song. From 2005 to 2014 it is 1 of 30, and the one is Amy Winehouse's Back to Black. I do not have a good theory for that stretch, and I am not going to invent one.

So they stay

Dropping them would cut 30 percent of the pool, including Thriller, Purple Rain and Back in Black, to close a hint that is right less than one time in five. The cases where it is right are mostly the ones where you already knew: a title track that was also a big single. Knowing that Back in Black was the hit and Pet Sounds was not is the kind of knowledge Expert is supposed to reward. If the name alone paid, I would feel differently. On average it pays under three points, and it costs you thirteen when the title track is a deep cut.

The one place I still see a small edge is the Daily. Par there comes from 5,000 simulated drafts of your ten albums, and those simulated players pick from the better half of each tracklist without reading titles. On a day when a title track is also the album's best song, a human who spots it gains a little on par that the simulation does not. It moves one pick out of ten. If the share grids start showing a pattern, I will write that up too.

Play Debutle