Reading the map

Every star is one athlete's whole career. Three things to look at:

By default you're seeing the top 125 from each sport plus the outliers, 0 stars total. If you want the full thing, check Show everybody and all 0 athletes in the data show up. It gets crowded.

The 3 axes

Quick idea first. Every season, every league has an elite tier, basically the best players that year (around the top 1 in 8 who played enough). All three axes measure you against that group. So a 1965 season gets compared to 1965, not to 2025. That's the only fair way I could find to put Bill Russell's era and Jokić's era on the same map.

Heads up: Height and Separation end up pretty close for most people, since the highest peaks usually come with the biggest gaps. Length is where careers really split apart.

The GOAT score

The GOAT score mixes the three axes. Right now it's Height %, Length %, Separation %.

I didn't just pick those numbers. I tried every combination in 5% steps and kept the one that lined up best with the all-time lists people already argue about: ESPN's #NBArank, The Athletic's NHL99 and NFL 100, Joe Posnanski's Baseball 100, FourFourTwo's top 100, and a few more. That mix is the Experts preset.

You don't have to agree with it. Drag the sliders under Your GOAT formula and watch the stars resize. The #1 at the top changes too.

Me? I'm a Longevity guy. Anybody can have a year. CR7 has 19 elite seasons, and Brady has 18. Flip to Longevity and LeBron takes #1, which is fair, I guess.

Where the model and I don't agree: Jordan 7th in the NBA. Kobe 24th. Maradona 24th in soccer. Take it up with the numbers.

Scores are scaled so 50 is the average of a sport's top 100 and 99 is the single best career on the map.

Comparing across sports

This was the hardest part. Every sport gets one "season value" number, using the best all-around stat I could get for it, and that number goes up against that season's elite.

Soccer was the worst to fit in, which is kind of ironic considering why I built this. The only stat I could find for every era was league goals.

The outliers

First, a problem I ran into. Height, Length and Separation move together. If you were great at your peak you usually lasted and usually lapped people too, so on the default map almost everybody sits on one diagonal from bottom-left to top-right. It's true, but it hides the interesting stuff.

So there's a second layout called Style view. Height stays on the X axis. The other two become "compared to your peak":

• Lasted: elite seasons beyond what your Height predicts. +5 means five more than players at your level usually get.
• Pulled away: Separation beyond what your Height predicts.

Zero is "about what you'd expect." Now the three axes barely overlap, and the weird careers end up at the edges where you can actually see them. Brady, Gordie Howe, Zlatan and Djokovic go straight up. Aaron Judge, SGA and George Mikan fall to the floor.

The GOAT score rewards players who were great at everything. That leaves out a certain kind of guy, the one who was elite at one thing and just okay at the rest. Those are some of the most interesting careers on here, so I tagged about 100 of them. Turn on Show the outliers and their names light up in the color of the axis they break. It works on any layout, but Style view is where they make the most sense.

• Built to last. Way more elite seasons than their score would tell you. Reggie Miller, Del Piero, Jim Thome, Big Ben.
• Burned bright. A peak that belongs with the greats, but not for long enough. Derrick Rose, Fernando Torres, Félix Hernández, SGA (for now).
• Ran away with it. One or two seasons where nobody was close. Willis Reed, Just Fontaine, Rod Laver.

How I picked them: in each sport, take the top 50 on one axis, then keep the 5 whose GOAT rank is furthest behind. Anyone who wasn't already in the default view got added.

What's missing

I'd rather tell you now than have you find it. NBA, NHL, MLB, F1 and tennis use every player in league history, so the elite tier is legit. NFL and soccer use a smaller group I put together, which is why their cards say lower data confidence.

• No playoffs. Clutch isn't in here yet. That's the next thing I'm building.
• Soccer only counts goals. That's rough on defenders, keepers and playmakers. Maldini deserves better.
• No women yet. The tennis data I used is men's only, so Serena isn't on here. That one's on me. A women's galaxy is on the list.
• Two-way players get split. Ohtani's hitting and pitching are scored separately, so he looks way worse than he is.
• Active players aren't done. Their Length can only go up, so they get a "Career in progress" tag.
• NFL stops at the 2024 season for now.

Data: Basketball Reference, the Lahman Baseball Database, the NHL Stats API, f1db, Jeff Sackmann's tennis data, Pro Football Reference, Wikipedia.

Controls

Why I made this

I started this during the World Cup, watching USA casuals go back and forth with football fans about who the real GOAT of GOATs was. Naturally I'm biased. CR7 is the greatest athlete of any sport (he is). But I wanted to actually back that up instead of just guess.

Most GOAT debates end up at stats anyway, so I leaned into that. The hard part was making the stats fair across sports that have nothing in common, which is what the whole elite-tier thing is about.

For the record, the model has CR7 19th overall and 3rd in soccer. I'm still working on that.

If you leave without being completely pissed off about where your guy landed, or who's missing, I'll call that a win. I love sports, I love numbers, and I like getting into people's heads. This is all three.

Babatunde Gbadegesin, Technology and Information Design at the University of Maryland. Data pipeline in Python, galaxy in three.js.