To all aspiring scholars,
I’ve noticed many of you are curious about the progress of our playtest version. Lately, we’ve been fully focused on optimization and bug fixes, and today I’d like to share a glimpse of our daily development life.
Yesterday, while passing by our artist’s desk, I was stunned to see a character wearing black stockings on their screen—I nearly dropped my jaw.
Can our Warring States era really have black stockings? Really?
The artist wiped the drool off their face (just kidding) and explained that the stockings disappeared after adjusting the lighting. They’re considering adding it as a visual toggle in the settings.
That reminded me: our dialogue interface has had a similar lighting issue all along.
So I assigned an urgent task to readjust the lighting—and finally, we can see characters’ faces clearly in dialogues now.
But beyond these visible bugs, the biggest challenge lies in optimization.
Although screenshots might suggest the system can handle 1,000 soldiers on screen at 90-120 FPS…
The originally designed AI was too realistic: each soldier calculates attack paths and actions based on distance and status, even including behaviors like fleeing or ganging up. But when they all interact simultaneously, the CPU simply can’t keep up.
The calculations consume massive resources, leading to high CPU usage—when computation can’t match the refresh rate, the game stutters noticeably.
Besides combat, overworld optimization is another major hurdle.
After capturing Wuyang County, I chose to plunder it—leaving the city starved and depleted. Merchants rushed to neighboring counties for goods, causing prices to skyrocket. These wealthy merchants then attracted bandits, and after local rebels were wiped out, the common people ended up worse off than before, with caravans frantically fleeing in all directions.
If this happened during a full-scale war between two states, hundreds of caravans and thousands of bandits could spawn at once, spreading across the entire map. Just imagining it makes me smell something burning… and it’s not my lunch.
The most effective optimization method? Cutting corners.
Soldiers off-screen won’t use full AI—they’ll simply charge forward and retreat when injured.
When a city runs low on supplies, goods will be magically “imported” instead of simulating real logistics.
The number of movable units on the overworld will be limited to two counties at a time.
We’ve invested heavily in enriching the freedom of this open-world experience, but to solve these issues in a short time, we’ve had to sacrifice some content that still needs polishing.
Even after aggressive optimization, we still have concerns about the current version’s stability and depth. We can’t in good conscience release it in this state. Therefore, we’ve decided to delay the playtest to November 15.
To those eagerly looking forward to playing during the National Day holiday—we sincerely apologize and hope for your understanding.
Our team will continue battling stubborn bugs throughout the holiday, ensuring the November build delivers more playable content.
— Xiao Xi & Qin Shi Huang