What Stick Wars is
Stick Wars is a Roblox stickman fighting game made by Devex Marketing’s own team. It went live on October 2, 2026, and we promoted it on Kash’s YouTube, TikTok and Instagram from the first day. It passed 3.4 million visits in its first five weeks.
We are publishing this for two reasons. Developers ask us whether we have built anything ourselves, and the honest answer is yes. And the project taught us a lot that is useful to anyone making a Roblox game, including a few things we got wrong. The rest of this page covers the plan, the content, the flops and the lessons. We keep to what our own notes support and do not publish revenue.
The game is inspired by stickman animation. Each round a player picks one of three owned stickmen and aims the opening launch with a direction and power setting. Server-side AI runs the fight on a two-dimensional board inside a three-dimensional lobby, best of three rounds. It is an auto-battler with player input at the pick and the aim. A game that decides in seconds who is winning is easy to understand in a short video, which was the point.
Design choices
Built to be watched, not just played
Early on, our priorities were set in a clear order: fight animations first, fight mechanics second, then the finisher scene. The stated reason was that fighting had to be polished enough to record clips. We added a finisher cinematic on the deciding knockout, flashy hit reactions, and a recording setup that could run hundreds of fixed fights on one board for trailer footage. If you make a game for short video, build the camera moments into the design.
Fast, readable matches
In our best stretch, rounds and matches were short enough for a phone player to say “one more.” Later we confirmed the instinct with data: on phones, long fights had noticeably lower continuation than short ones.
An easy first experience
The first three matches are against easy scripted bots, so new players get wins and learn the controls. After that, real players come first, with a bot fallback so nobody waits long. Players below a certain level are routed to a separate new-player place.
Kid-friendly economy
We started with crates. Our owner then decided he did not want a game with random-chance purchases aimed at younger players, and we replaced crates with a level-gated restock shop where everything is buyable and nothing is gambling. The details are in our Roblox game monetization guide.
Creator tools inside the game
Because our business is working with creators, we built tools for them: a green-screen mode, tracking, instant replay and custom matchups, with a whitelist of creators allowed to use them. The goal was to make it easy to film a good-looking fight, whoever was filming.
The content plan
Promotion ran across Kash’s YouTube, TikTok and Instagram, with the game’s own channel carrying the faceless, game-first posts. A few rules from our own playbook:
- A daily cap per account. No more than three posts a day per account, so no account turned into spam.
- Consistent branding. Every post carried the game name on screen, kept faceless to match the art style, and pointed players to a shared in-game code.
- Mobile-first framing. Action in the center of the frame so it survives a vertical crop.
- Versus format. Clips titled “Who wins? X vs Y” or “Does X beat Y?” show two stickmen fighting and invite a comment. We took the idea from a popular auto-battler Shorts channel and applied it to our roster.
- No duplicates across accounts. We never post the same clip to two channels on the same platform.
Our notes also include a playbook for the next stage, built from studying how other developers grew on short video. The common thread was not editing polish. It was a reason to show up: a fixed weekly event, a limited reward, a question for players, and the owner as a visible character. We recorded it as a recommendation, not something we had proven on Stick Wars, and the weekly-event version is a plan rather than a result. If you want the general version, read Roblox Shorts and Clips Marketing.
Reposting and what flopped
Not everything worked. On the game’s own YouTube channel, our early versus-style Shorts underperformed badly, and we have since set them to private.
We then planned a repost test: pick a set of those Shorts, post them again under a layout and music rule chosen up front, and log views at fixed intervals for each repost. As of this writing that test is on hold, waiting on decisions about copyrighted songs and a boxed layout. So we do not have repost results to share, and we will not pretend to.
What we take from it so far:
- A format that works for a big channel does not automatically work for a new one.
- Titles with a question are not enough if the first second does not show something surprising.
- Decide music and layout rules before reposting, so the test is clean and you do not risk copyright problems.
- Treat your first batch of posts as experiments. That is the same advice we give developers in our marketing guide.
The numbers we publish
The one number we put on this project is the public one: Stick Wars passed 3.4M+ visits in its first five weeks. We keep other internal figures private. What we can say qualitatively is that players arriving from Roblox’s Home recommendations stuck around far better than players from sponsored ads in our first ad test. That is one game over one window, not a rule about ads, but it matches what Roblox has said about recommendations favoring retention. For a fair look at the trade-offs, see Roblox Ads vs Paying YouTubers.
The update that hurt
For a stretch in October the game was a fast, flashy auto-battler with no levels and one automatic ability per fighter. Then on October 26 we shipped a major rework in a single publish. Many big systems changed together: stickman levels and upgrade costs, a new economy, rebuilt character kits, buttons the player had to press, much longer matches, damage-only basic hits and a new onboarding.
The result was the opposite of what we hoped. Concurrent players fell sharply that night and stayed lower, and the share of new players continuing from match to match dropped. Many PvP matches had a big level gap, and the stronger side won the large majority of them. Rarity also did not mean strength, which made the economy feel unfair. Our owner’s own words on October 29 were that the game was young, he was very ambitious, and the huge update kind of ruined the game concept.
The fix was to stop shipping everything at once. We agreed four releases with one theme each:
- Feel: automatic play on by default, shorter matches, rounded rewards and a fix for failed save loads.
- Progression: levels become mastery with no stat power, plus titles and borders.
- Balance: target win-rate bands per rarity set from live data.
- Return reasons: a battle pass and a trophy road.
We also changed process: A/B tests for anything that alters play or progression, version tags on releases, and quiet windows so the data stays readable.
Lessons for your game
- Know what players love before you change it. Ours loved a low-effort, watchable, fair fighter. We added effort and grind.
- One theme per update. You cannot read, or undo, many changes at once.
- Test before you roll out. Give half of new players the change first.
- Avoid stat power in fair PvP. It reads as pay-to-win.
- Phones are the audience. Shorter matches and simpler controls win.
- Retention over traffic. Roblox ranks on multi-day return, and promotion cannot fix a leaky first week.
- Show round numbers. Small, odd rewards feel stingy.
- Never let a failed save look like a new player. Retry, then tell the player their progress is safe.
- Treat early posts as tests. Most of ours underperformed. What matters is what you learn from them.
The design-side version of these lessons is in How to Make a Roblox Game That Gets Players.
What this does not prove
We would rather be clear than impressive. This is one game. We built it and promoted it on Kash’s channels from day one, and it passed 3.4 million visits in five weeks. We did not run a controlled experiment, so we cannot say how many visits came from each channel or what would have happened without the posts. Visits are also not players or revenue, and we are not publishing revenue. Your results will depend on your game, your page and your first minute.
If you are weighing options, a small test with tracking tells you more than any case study, ours included.
Frequently asked questions
How many visits did Stick Wars get?
Stick Wars passed 3.4 million visits in its first five weeks after launching in October 2026. The game was built by Devex and promoted across Kash’s YouTube, TikTok and Instagram from day one.
Did the promotion cause the 3.4M visits?
We cannot say that. We built the game and promoted it from day one, and visits also come from Roblox recommendations, search, friends and ads. We did not run a controlled test, so we describe what we did, not a proven cause.
What kind of game is Stick Wars?
It is a stickman fighting game with an auto-battler feel. Players pick one of three stickmen, aim an opening launch, and watch server-side AI run the fight, best of three rounds. It was inspired by stickman animation and designed so fights look good in short clips.
What went wrong?
We shipped one large update that changed many systems at once, including levels, the economy and match length. Player numbers and our new-player funnel got worse. We wrote a postmortem and split the fix into smaller releases.
Can Devex promote my game the same way?
We sell guaranteed views on the posts we make about your game across Kash’s channels. How many viewers become players depends on your game, your page and your first minute. Stick Wars is one case, not a promise about yours.
Where can I learn how to design a game for clips?
Start with our guide to making a Roblox game that gets players, then read the Shorts and clips marketing guide. Both use lessons from this project.