The scores were never the point

So The Sports Machine is, in a way, the third act of something that started with a kid and a pencil. First I copied the data by hand. Then I taught a computer to fetch it for me. Now I'm building something that gathers it, updates it, and makes sense of it on its own. Same love, three different amounts of leverage. It's a homecoming — the subject that taught me to build, finally getting my best work.

The through-line is almost embarrassing in how straight it runs. The kid copying standings became someone who taught himself to program on exactly this material, and then someone who builds systems that do the copying for him. Every rung on that ladder was the same instinct at a different amount of leverage: take the thing I care about and hold onto it more permanently than the tools around me allowed.

Spending that much of my life inside sports data — across just about every site that serves it — you start to notice something. The data itself is a commodity. The final score of last night's game is the same number everywhere. ESPN has it, a hundred smaller sites have it, and none of them has a better version, because there's no such thing as a better score. It's just the score. In sports, the facts are free and identical everywhere. That sounds like a problem for anyone wanting to build here. I've come to think it's the opposite.

Because if the data is a commodity, then the data was never the product. The product is everything around it — how fast it loads, how it looks, how easily you find the thing you came for, what it helps you notice that you wouldn't have on your own. The scores are table stakes. The experience is the whole game.

I've said a version of this so many times it's become a kind of creed: if everyone has the same numbers, the numbers can't be the moat. Everything worth defending gets built on top of that commodity, and that part takes years.

That's what I set out to build: the sports site I always wished existed. Clean, fast, modern — all the sports I care about in one place, pro and college, football through hockey, with the numbers I love front and center and the design finally working for me. I wanted looking at sports data to feel as good as I'd always thought it should, back when it was just me and the sports page.

Every existing site made a trade I didn't want to make. The big ones cover everything but bury the score I came for under three ad units and a video that autoplays. The clean ones cover one league beautifully and abandon the rest. Nobody had decided that breadth and care were allowed to coexist — that you could have all the sports and have it feel light. I wasn't trying to out-fact anyone; I was trying to make the same facts feel like something.

The name points at the other half of it, and it doesn't mean what people assume. “Machine” isn't about power. It's about a platform that runs itself — one that updates its own data, refreshes its own scores, and generates its own analysis without anyone sitting behind it typing. In a way it's the same thing I did at seven, just handed off: the gathering and the updating happen on their own now. That automation isn't a feature bolted on the side. It's the reason a project like this can exist at all.

Because covering all of sports, accurately and constantly, is normally the job of a newsroom. What I have instead is the thing sports data taught me in the first place — how to turn a messy, ever-changing pile of numbers into something structured and self-updating. So staying comprehensive and current happens on its own. Most of what makes The Sports Machine run is invisible: quiet plumbing underneath, doing the work a building full of people would otherwise do.

It's worth being concrete about what "runs itself" means. Underneath the scores is a loop with three jobs: ingest the raw feeds, clean and reconcile them, and serve the result as structured data anything can read. During a live game those scores refresh roughly every fifteen seconds, and that number reaches back through the pipeline. The ingest has to keep polling, the clean stage has to reconcile fast enough to not fall behind live play, and the serve layer has to hand out fresh state while every viewer refreshes at once. Real-time isn't a moment you hit; it's a treadmill you keep running every game day.

The scale is the part that still catches me. The warehouse underneath tracks north of 630,000 games and 25,000 teams across twelve sports, updating live. None of that is a person typing — it's the seven-year-old's habit handed off in full, at a size no ledger of mine could hold.

The half of this that no visitor will ever notice is where almost all the real work is. Raw sports data arrives messy — duplicates, gaps, and, worst of all, the same team showing up under two different IDs from two sources. Left alone, one of those identity collisions quietly corrupts every standings table and every historical join that touches it. So most of the machine isn't the scoreboard at all. It's the reconciling — the patient checking of one source against another until the numbers add up and I believe them.

This is the unglamorous half, and I chose it on purpose. The glamorous half is the clean chart that makes something obvious; the unglamorous half is everything that has to happen first so the chart isn't quietly lying to you. When the facts are free and identical everywhere, the only advantage that lasts is that mine are trustworthy — and trust is earned in the boring, invisible work nobody wants to do, which is precisely why it's defensible.

Put it together and you get what I actually believe about this category, and maybe a few others: when the data is free and everyone has it, you don't compete on the data. You compete on everything else — the design people feel but can't name, the breadth the big sites cover thinly, the automation that keeps it all fresh. The facts are a commodity. The feel isn't, and the feel is where there's still room to build something better.

I'll be honest about the scale of it. I'm not out-reporting ESPN on breaking news, and I'm not trying to. Part of what The Sports Machine is, is a proving ground — a place to build a genuinely modern, self-running platform and see how far it can go. And part of it is a real bet: that even in a space full of giants, there's room for something built with more care, covering the corners they cover thinly. Both are true at once.

It's a business bet too — I'd like it to work as one. But in the order I hold for everything: make it genuinely good and worth coming back to first, and let the rest follow. That's less a strategy than a temperament.

Mostly, though, it comes back to where it started: a kid at the table copying standings out of the paper because he loved them too much to just read them once. The Sports Machine is my grown-up version of that same impulse — not because the world was short on scores, but because I've loved this stuff my whole life and wanted, finally, to give it a home worthy of it.