TL;DR: Close tabs, disable heavy extensions, enable hardware acceleration, use wired network, lower zoom to 100%, update GPU drivers.
Tab hoarding cost
Each tab competes RAM and GPU compositor—close Spotify video background tabs.
Extension audit
Incognito window extensions disabled test—if smooth incognito culprit extension identified.
Hardware acceleration toggle
Chrome settings system—on for GPU decode usually helps canvas games off rare laptop bug try flip once.
Network stability
Game mostly local JS; lag spikes from CPU not bandwidth unless assets re-fetch each run broken cache.
GPU driver age
Intel integrated update quarterly—stutter on canvas 2D surprisingly driver sensitive old machines.
Power plan high performance
Windows battery saver caps CPU—plug in laptop tournaments.
Display scaling
125% Windows scaling adds render step—native 100% test PB environment.
Kill overlays
Xbox Game Bar, Discord overlay, MSI Afterburner OSD—disable during timing tests.
Benchmark before after
Same RLGL segment record frame feel subjectively or FPS overlay if permitted.
Maintain clean profile
Dedicated Chrome user profile games only—minimal extensions bookmark game URL.
Dedicated Chrome profile games
Minimal extensions bookmark official URL—clean environment.
Kill Xbox Game Bar overlay
Overlays add input lag feel—disable during timing test.
High performance power plan
Battery saver caps CPU—plug in laptop tournaments.
100% scaling not 125% Windows
Render step adds latency marginally—test PB environment.
Incognito vs extension binary search
Systematic find lag culprit before buying new hardware.
Incognito window as the extension-off test
An incognito or private window often disables extensions. Play three attempts there. If they are smoother than the normal window, an extension in the normal window is costing you frames. You do not have to live in incognito forever. You have to know. Once you know, disable the heavy extension in the normal window for practice and keep it for shopping if you must. The three attempts are the experiment. Write “private window smoother” or “no difference” so you do not repeat the experiment every day.
Background video closed before the first green
A video playing in another tab, even paused on some browsers, shares the machine with the doll. Close it. Music players that are only audio are usually fine. A video of a concert in a background tab is not. The test is cruel and simple: if the fan spins up when the video is open and calms when it is closed, the video was doing work. The doll needs that work more than the concert does. The concert can resume after the session.
Zoom at 100 percent so the lamp is not a blur
Browser zoom at 150 percent makes the doll large and the page scrollable. A scrollable page steals arrow keys. Set zoom to 100 percent, then fullscreen if you want size. If you need larger type for the menu, enlarge it and then return to 100 before Start, or use the operating system’s scale rather than the browser’s zoom. The arrow keys should move the token, not the document. Test one arrow press on the menu. If the page scrolls, you are not ready to start.
Hardware acceleration toggle tried once, not every night
If the game stutters, toggle hardware acceleration once, restart the browser, and play three attempts. If nothing changes, put the toggle back and look elsewhere: heat, extensions, wireless display. Toggling it every night adds a variable you will forget. One experiment, a note, and a decision. “Acceleration on, still stuttered, looked at heat instead” is a complete note. A week of random toggles is not a diagnosis.
Private-window result written as smoother or not
Write one of two phrases after the three private-window attempts: “smoother” or “no difference.” Do not write a paragraph. The phrase decides whether you spend time on extensions. “Smoother” means disable one extension tomorrow and test again. “No difference” means stop blaming extensions and look at heat or the wireless display. A paragraph will let you do neither. The phrase is a switch. Flip it.
Zoom test with a single arrow press
At 100 percent zoom, press one arrow while focused on the canvas and watch whether the page moves. If the page moves, you are not on the canvas. Click it and press again. If the page still moves, something else is bound to the arrow. Fix the binding or play with the mouse for this session and write “arrows stolen.” Do not fight a scrolling page for ten attempts and call it practice. The page is practicing against you. Win that fight once, then play.
Video tab closed even when it is paused
A paused video can still hold a decoder. Close the tab. If the fan was loud with the pause and quiet after the close, you have the answer. Play the next attempt only after the fan drops. Logging a best while the machine is still hot from a concert video measures the concert. Close it, wait for the fan, then start the attempt you intend to keep.
Extension named on the day you disable it
When “smoother” sends you back to disable an extension, write the extension’s name in the note. “Smoother” alone is a mystery you will re-solve. The name is the solution. Disable that one, play three attempts, and write “better” or “same.” If it is the same, turn it back on and pick a different extension tomorrow. One per day. A day that disables five teaches you nothing you can repeat. A day that names one teaches you a setting.
