A half-year check-in on a hobby project that got a little out of hand.
Back in February I published a piece called Why I Built SaltyTrack — the origin story. Knockoff Xbox controller, my own fledgling MUGEN setup, rocking my Shin Akuma main, rediscovered SaltyBet late one night, got tired of betting blind on 10,000 fighters I couldn’t possibly memorize, and decided to build the tool I wished existed. Just help me make an informed decision, man!
That was written about a month into the project, when everything was still shiny and I was still doing my He-Man impression every time a feature worked. I used to do that. I still do, but I used to, too.
It’s now been a little over five months since launch. SaltyTrack has quietly grown from “a scrappy thing that recorded match results” into a data brain that has tracked north of 620,000 SaltyBet matches, serves live predictions right on the stream, and publishes more than 9,000 crawlable fighter pages.[1] And I’ve learned a ton — most of it the kind of stuff you can only learn after you launch, not before.
So this is the sequel. Less “here’s why I started,” more “here’s what half a year in the salt mines actually taught me.” Some of it is technical. Most of it isn’t.
When I started, I thought I was building something simple — something to help you know which way to bet, something slightly better than “well, I kind of recognize that name.” I’d be lying if I said I knew at the outset you’d get all the juicy info you get now: win rates, matchup history, the tidy little overlay.
What it became is something I genuinely didn’t plan for: a prediction engine that now calls fights correctly more often than the betting crowd does (LET’S GO!!), a growing library of data journalism about a Twitch meme, and a Chrome extension that puts all of it one glance away while you bet. The predictions started as glorified rules of thumb and, over the spring, sharpened into something that actually beats the room on a typical night.[1] I did not expect the “for funzies” data project to end up good. Real good, if I do say so myself.
It’s hard to pick just one thing that surprised me the most. I was, and still am, surprised at how much there is to do when it comes to improving and building the platform. I’m amazed at how much I’ve learned from building SaltyTrack, and how much I enjoy working on it.
I’ve been humbled and amazed at how the platform’s usage has grown — we have a modest but extremely loyal following, without any paid promotion or user acquisition.
I’ll say it plainly: I underestimated just how powerful the draw of building something from scratch would be — taking an idea that only existed in my head and making it real. How incredibly seductive and consuming it would become. I’ve lost many a weekend just iterating, improving, and otherwise refining the experience — a lot of it invisible to the everyday users of the extension and the website.
Here’s the plot twist nobody warns the hobbyist coder about: writing the first version is the easy part. Keeping a thing alive is the hard part.
A few lessons that cost me something:
The hardest moments were definitely the beginning stages, and I think that’s in line with how the first time you do anything is going to be the hardest, most difficult stretch — because you don’t know what you don’t know.
I was humbled more than once implementing the secret sauce that makes SaltyTrack accurate and good; keeping that secret sauce fresh, tasty, and appealing to everyone who visits is an ongoing, continual process.
Somewhere in there I also got a crash course in the necessity of certain elements of cybersecurity — the unglamorous, keep-the-lights-on side of running anything that lives on the public internet. It turns out that keeping a service healthy, resilient, and available is its own whole discipline, and picking it up was one of the more unexpected side quests of this whole project. So I got that going for me!
I’m proud of SaltyTrack. I’m also very aware of the scar tissue.
Honestly, this is the hard part, because when I hear the phrase “what would you do differently,” to me it implies that the original saga is over — and to me, we’re very much still ongoing. There were some aspects of the secret sauce I’d have tuned differently from the jump, but I think those are things I’d only have known to tinker with after some iteration. Like getting the right blend of spices in a meal, it takes some time, some adjustment, and some data to go off of.
I also think I would have created more content — more insights articles — from the very beginning. MUGEN is a niche within a niche, and part of what I want to accomplish with this site is simply making more people aware of MUGEN in general. Content about “What is it? How does it work? Why is it fun?” helps get more people involved in the community.
The honest meta-lesson underneath all of these: AI coding tools got me from zero to a working product, but they didn’t let me skip understanding what I shipped. The gap between “it runs” and “it runs correctly, forever, unattended” is exactly the gap you have to learn to see — and no tool sees it for you. That’s still the most valuable thing this project taught me.
Five months in, the nostalgia that started this whole thing hasn’t worn off. If anything it’s deeper now, because I understand the history of MUGEN and the ridiculous, wonderful machine underneath SaltyBet better than I ever did as a kid downloading characters onto a laptop.
And like I said, this is a project I continue to manage and oversee. It very much receives active development. While life takes up portions of the time I’d otherwise dedicate to it, I have no plans to sunset or stop working on it any time soon. I make sure to attend to it in some form or fashion throughout the week.
Part of that is because I so thoroughly enjoy the building, the creating, the changing, the tinkering, the implementation. I derive a great deal of personal satisfaction and pride from putting something into the universe that others find valuable, useful, and fun to use.
I love the craft of MUGEN. I love the unique works, activities, behaviors, and tones you find in the community.
And since I was exposed to so many of these characters and franchises as a young boy, I think part of my boyish love of sci-fi, fighting games, video games, drawing, and computers is part of what compels me to keep refining and working on the craft. I loved the Mortal Kombat movie as a kid. Same with the Street Fighter movie, the video games, and the franchises.
I’ve owned a lot of video game consoles in my time, and I find video games in general to be an avenue of reflection, growth, and personal enjoyment. I think you can learn and take something from the games you play over time. Some are pure surface-level fun; but some are deep — they hit hard with life lessons and life truths, and the community will know exactly what I’m talking about.
I look forward to continuing to build and refine this platform, this product, this service. If I could somehow get all of our prediction accuracy up into the mid-to-high 80s, that would be an amazing milestone for me. It’s perhaps always a goal out on the horizon — given the sheer variability in what an individual creator can make a MUGEN fighter do — but that won’t stop me from trying!
For the fellow builders and the curious, the short version of the journey so far:
| When | Milestone |
|---|---|
| Jan 2026 | First commit. A scrappy collector that could read SaltyBet and record match results. |
| Feb 2026 | Public launch, the insights page, and “Why I Built SaltyTrack.” |
| Spring 2026 | The prediction model matured from rules of thumb into something that beats the crowd on a normal night. Meta reports, tier deep-dives, and franchise breakdowns started shipping. |
| Summer 2026 | More than 9,000 individual fighter pages went live, and the Chrome extension kept getting smarter about putting the data where you’re actually looking. |
| Today | Still running. Still tracking every match. Still for funzies. :) |
If you’ve read this far, you clearly love this dumb, beautiful game as much as I do — so let me do the honest plug I always do. Everything I’ve built lives one click away in the SaltyTrack Chrome extension: live predictions, win rates, and head-to-head history overlaid right on the SaltyBet page, so you never have to bet completely blind again. It’s free, it’s a labor of love, and it’s the whole reason this project exists. Give it a spin, and I’ll see you in the mines. Salt on. 🥒
Predictions are for entertainment. Bet responsibly — with Salty Bucks, which are gloriously fake. SaltyBet uses virtual currency only. No real money is wagered or exchanged. SaltyTrack is not affiliated with SaltyBet.