From the Clubhouse
What 10,000 Simmed Tournaments Taught Me About Building a Golf Game
Quick Start, mobile players, ridiculous bugs, and why caring about your golfer is harder than getting people to click play.
Before I started messing around with prediction models and some of the newer stuff, Rainy Day Golf was basically just a golf career sim.
And building that first version taught me a lot, mostly because people immediately started finding all the things I had gotten wrong.
Five minutes was way too long
The original setup process took about five minutes before you could actually play a tournament.
At the time, I didn't think that was that bad. You created your golfer, made a few choices, got everything set up, then started your career.
People hated it.
One of the first pieces of feedback I got was basically: just let me play.
So that night I added Quick Start. Instead of spending five minutes setting everything up, you could be playing in around 30 seconds.
From the career-sim launch story in this post: full setup versus Quick Start.
Then the usage jumped.
Within four days, the game had gone through more than 10,000 simulated tournament rounds. About 300 people had played, which worked out to roughly 33 rounds per person.
- 10,000+Simmed rounds
- ~300Players
- ~33Rounds / person
Early career-sim traffic after Quick Start, as described in this post.
That was probably the first time I realized how much a small amount of friction matters in a browser game.
If somebody clicks a random golf game from Reddit, they're not giving you five minutes to convince them it's fun.
They're giving you about 30 seconds.
Apparently I was building a mobile game
I also built the whole thing mostly on desktop.
Then I looked at the analytics.
Depending on the week, somewhere around 72–80% of people were playing on their phones.
- Phone 76%
- Desktop 24%
Midpoint of the early career-sim analytics range in this post (~72–80% mobile depending on the week).
I hadn't planned for that.
The game happened to work decently on mobile, but it definitely wasn't designed mobile-first. I was sitting at my desk tweaking desktop layouts while most people were apparently playing it on a phone.
That changed how I started thinking about basically everything afterward.
Menus have to be faster. Text has to be shorter. Buttons can't be tiny. You can't assume somebody wants to sit there and study a giant stats screen.
A lot of people are probably opening the game for a few minutes, playing a round, then closing it again.
People found some ridiculous bugs
This was probably my favorite part of putting it online.
You can test something 100 times yourself and somehow miss the dumbest stuff.
One person pointed out that finishing 102nd showed up as 102th.
Easy fix.
Another person found a much better bug: one tournament never actually ended after 18 holes.
You could just keep playing.
And because the leaderboard kept updating, you could technically grind your way from wherever you finished all the way back into first place.
I had accidentally invented the world's longest golf tournament.
That got an actual end-of-round screen pretty quickly.
Then a moderator from the Basketball GM community spent a bunch of time digging through the career mode and found problems I probably wouldn't have caught for a while.
Players weren't aging correctly over long careers.
You couldn't see your current attributes while deciding what to train.
And my favorite: a rookie golfer could somehow have ten years of career analysis talking about him like he'd been on tour forever.
Those were the kinds of things that made me realize people were actually paying attention.
They weren't just clicking around for two minutes and leaving. Some of them were simming years into a career and trying to understand how the game worked.
The feedback I keep thinking about
The most useful comment wasn't actually about a bug.
Someone basically told me there wasn't enough reason to care about your golfer early in the game.
You could play tournaments. You could climb the leaderboard. You could improve your stats.
But why should I care whether this particular fake golfer succeeds?
That stuck with me.
Things like rivalries, milestones, career goals, records, personality, whatever else makes a sports career feel like a story probably can't show up 20 seasons later.
You need at least a hint of that almost immediately.
It's something I'm still trying to figure out with Rainy Day Golf.
Getting someone to start playing turned out to be a pretty solvable problem. Remove the setup, make it fast, and let them do the fun thing.
Getting them to care enough to come back tomorrow is a lot harder.
The 10,000 rounds were awesome to see.
But the number I'm probably more interested in now is how many people wanted to play round 10.