Did Cutting Mode 7 Make the Game Faster?
You voted Mode 7 off the idea board, and we deleted it. Did that make the game faster? Here is what we measured after the cut, and where the data stops.
The mode was the co-op escort mode we shipped in June. It got 412 votes to cut it. The next closest idea got 189. We ran the numbers and shipped the deletion in four hours. The assets went with it.
Nobody played it enough to justify keeping it alive. Out of 1,204 sessions last month, 11 touched Mode 7. Eleven. Every time you queued for a match, the matchmaker also checked for Mode 7 players, and that check cost about 80 milliseconds per queue. Not much on its own. But it stacked with everything else, and you felt it.
Every mode we add is a cost we pay for as long as it exists. It shows up in load time and matchmaking. It shows up in hours spent debugging a mode nobody plays. If you want to play online games without downloading anything, somebody still has to decide what stays. You decided. We just listened.
Here is what happened to the servers. Average match load dropped from 2.1 seconds to 1.2 seconds. Frame time in DOMINION dropped from 28 milliseconds to 16. That is about 43 percent less frame time. We did not touch the renderer or rewrite the netcode. We removed code that was running in the background for zero players.
We want to be straight about one thing. We designed Mode 7. We thought it was good, and we had a whole roadmap slide about it. You played it and said cut it. That stings a little. It is also why this studio works the way it does.
The Vote Count That Made the Call
The idea board is not a suggestion box. A vote count beats a roadmap we wrote in a doc at midnight. Mode 7 got cut because the vote was not close. We posted the proposal on a Monday. By Tuesday it had 97 votes. By Friday it had 412. We had set a one-week window, and the community cleared it in four days.
Here is the part that surprised us. The people voting to cut Mode 7 were not angry. They were practical. The top comment said they would rather have RUSH maps than an escort mode they never touched. That comment got more upvotes than the proposal itself. You were not voting against the mode. You were voting for your time.
We track retention per mode, because retention shows what reviews often miss. Mode 7 had 4 percent day-two retention. RUSH had 31 percent. DOMINION had 24 percent. Mean session time in Mode 7 was 3 minutes and 12 seconds. In RUSH it was 18 minutes. Those numbers made the argument. The vote confirmed what the data already showed.
We once thought cutting something meant admitting failure. A small student team in Tunisia cannot afford that kind of ego. We have about 30 hours of dev time a week between classes. Every hour spent maintaining a dead mode is an hour not spent fixing spawn points or making your shots land. You gave us 30 hours back per month. That is the real prize.
If you pitch an idea next month, pitch a cut. Seriously. Some of the fastest wins on the board are deletions. Pitches that start with remove, for a thing that annoys you, pick up votes faster than most add ideas. You already proved that.
What You Actually Feel in a Match
Frame time is what your hand feels, not what a spec sheet says. At 28 milliseconds per frame you get about 36 frames per second. At 16 milliseconds you get about 60. That is the jump from playable to good in a browser shooter. On a school laptop. On classroom WiFi. No install, no credit card, no launcher.
Matchmaking got faster too. Average queue time dropped from 14 seconds to 9 seconds. The matchmaker no longer checks five mode buckets. It checks four. Less branching means the server picks your match sooner. In a 16-agent room, the last slot was the slowest to fill. On a normal day it fills in under 6 seconds now.
The biggest gain was in DOMINION, our RTS mode and the heaviest thing we run. Every frame, it checked whether any Mode 7 escort objectives were still alive in memory. After the deletion, DOMINION held close to 60 frames per second on hardware that had been choking at 38. One player wrote to say it felt like a different game. It is the same game with 43 percent less baggage.
The load screen got shorter too. The first match hits the loading bar in 30 seconds on a cold cache, and in about 11 seconds on a warm one. We cut 1.4 megabytes of assets that only Mode 7 touched. That sounds small until you remember some of you play on phones at a bus stop. Every megabyte is a second of your life.
And here is the part we did not expect. Crash reports dropped. We averaged 9 crash reports a day tied to Mode 7's escort pathing. That is now zero. You did not just make the game faster. You made it stop falling over.
The Tradeoff We Accepted
Cutting Mode 7 cost us retention, and we have to own that. Overall day-two retention fell from 22 percent to 19 percent. Three points. Some of you only logged in for Mode 7, and you have not come back. The numbers dipped for five days. We did not panic, and we did not revert. Three points down for a game that runs faster is a trade we would make again. We will not pretend it was free.
We also lost the co-op players. Mode 7 was the only mode built around escorting a slow objective together. RUSH is wave defense. It is intense, but it is not the same calm teamwork. We heard that. We do not have a replacement yet. If you want one, put it on the board and get the votes. That is how this works, and you know it by now.
Then came the bugs we created. Deleting a mode is not like deleting a file. Mode 7 shared spawn logic with two campaign missions. We shipped the cut on a Friday. On Monday we spent four hours patching two missions that spawned you outside the map. Twenty-four hours to ship the fix. Not our proudest patch. You reported it fast, and your reports fixed it.
The lesson here is that cuts are features too. Most studios talk about what they add. Few talk about what they remove. The board pushed us to be one of the ones that does. If you are judging any indie studio, ask them what they cut last. If they cannot answer, they are probably carrying dead code.
We kept the Mode 7 art in the Classic build. The 2020 Unity build is still online and untouched. If you want to see what we killed, you can walk around it there. It runs like a museum piece, slow and wobbling. That is why it is a museum now.
The Next Vote Is Yours
The idea board is open right now. Everyone gets one vote per idea. That includes you, and it includes the person who has never played before. No exceptions, no cash shortcut. The two in-game currencies are earned by playing, not bought, and the same rule applies to your say in the roadmap.
Here is what sits at the top of the board today. A request to make RUSH drop better loot is at 201 votes. A DOMINION map is at 174. A Loadout Lab capture mode is at 88. None of these are locked in. Top-voted ideas go Planned, then Building, then Shipped, and you can watch each one move.
We will not tell you what to vote for. That would defeat the point. We will tell you what Mode 7 taught us. Cuts move fast. Modes take longer. We built Mode 7 in three weeks and cut it in four hours. If you want speed, vote for the thing that takes friction out of a game you already love.
The 412 of you who voted to kill Mode 7 did something most studios never let players do. You looked at work we were proud of and said it costs more than it gives. You were right. That is not a small thing for a group of students building a game between lectures. You are not just playing our game. You are running part of the studio.
Your call on Mode 7 is why the numbers above moved. If you have not played since the cut, queue one match. Feel the 9-second queue. Feel the 16-millisecond frame time. Then go vote on the next idea, because the next speedup is sitting there waiting for somebody to call it.