You drop the second star into row 3 and your eyes jump straight to that interesting-looking region in the corner. Ninety seconds later you're back in row 3, hovering over column 5, working out whether a star could fit there. It can't. The row was finished a minute and a half ago. You just paid for the same deduction twice.
That's where solve time actually goes. Not the hard deductions. The re-checking of things you already knew. Getting faster at Star Battle mostly isn't about cleverer techniques. It's about three habits that stop the leaks.
Three, not ten. An earlier version of this article listed ten tips. When we reread it, three of them turned out to be the same elimination idea with different names on it, and one was "practice daily." Nobody gets faster reading that. So we cut it down to what's real.
Habit One: Scan in the Same Order Every Time
Fast solvers don't dart around the grid hunting for something interesting. They run a loop. Ours, for what it's worth: smallest regions first, then rows top to bottom, then columns left to right, then back to the regions. The specific order barely matters. Pick any loop and you'll do fine. What matters is that it's fixed.
Here's why. When you scan in a fixed order, "there's nothing left in the rows right now" is a fact. When you wander, it's a feeling. And you will keep going back to re-verify feelings, which is exactly the row-3 tax from the opening.
A fixed loop also closes things out the instant they finish, and that's where the compounding starts. Take a 9×9 board, two stars per row, column, and region. Row 3 already has a star at column 1; a second one just landed at column 7. Row 3 is done, and in the loop a done row gets X'd out on the spot — every remaining open cell in it.
Now suppose one of those Xs lands on row 3, column 4, the last row-3 cell of the region sitting above it. That region's open cells are suddenly row 2, columns 3, 4, and 5. All in one row, so both of its stars go in row 2, and two stars in three consecutive cells can only be the non-touching pair: columns 3 and 5. Two stars placed and row 2 finished, all hanging off one routine X.
None of that came from a clever read of the board. It came from closing out a finished row instead of letting your eyes wander off. If the confinement step is new to you, the region confinement article covers it properly. Of everything we teach, that's the one to know cold.
Habit Two: X Like It's Part of the Move
A star isn't placed until its consequences are on the board. The moment a star lands, X all eight neighbors. The moment a row, column, or region gets its second star, X out whatever's left of it. Before your eyes move on. Not "in a second," not mentally.
Every unmarked X gets paid for twice: once when you work out the cell is dead, and again later when the naked square pulls your eyes back in and you work it out again. Multiply that by twenty cells a puzzle and you've found several minutes that no technique was going to give you. The logic behind all these Xs is the elimination technique, worth one read even if you're sure you already do it.
Our play screen has an Auto-mark Xs setting, on by default, that handles the mechanical part: eight neighbors when a star lands, automatic close-outs when a row, column, or region fills. Honest advice: turn it off for a week. Doing the Xs by hand is what teaches your scan what a dead cell looks like, and that recognition is the actual skill. Once the recognition is there, turn the setting back on and spend your attention on deductions that deserve it.
Habit Three: When the Scan Stalls, Slow Down on Purpose
Two clean passes through the loop with nothing to show means the puzzle will not yield to a third, faster pass. This is the point where most solvers speed up, rescanning harder, as if the deduction is hiding behind their own eyelids. It's exactly backwards.
Stalling means the board needs a different gear, not a faster one. Stop scanning. Pick two or three adjacent rows and count how many stars their regions can still supply. Check each region for confinement. This feels slow. It's slower per minute and faster per solve, because one counting deduction routinely breaks a board that thirty seconds of frantic rescanning wouldn't dent. The illustrated strategy guide lays out all nine core moves in the order you should reach for them.
About the timer: the play screen shows one, it pauses when you leave the tab, and you can hide it in Settings. Hide it. Not forever, just until the X habit runs without thinking. Racing a visible clock is how "X the neighbors" turns into "X the neighbors later," and one sloppy star costs more than the clock ever gives back: you'll undo it three deductions downstream, along with every X you hung off it.
Where the Gains Stop
Fair warning: these habits are worth minutes, and everything past them is worth seconds. Once you're scanning in a loop, X-ing as part of the move, and downshifting when you stall, further speed comes from pattern recognition: glancing at a clump of cells and knowing its star count without counting. That's built by volume, not by tips. There's no fourth habit hiding in someone else's listicle, ours included.
So the last piece is just reps. The daily medium puzzle is the right treadmill for this: hard enough that the habits earn their keep, light enough to finish every day. And if you want the scan-order habit stress-tested somewhere new, play Queens — one queen per row, column, and color region, same no-touching rule. The loop transfers directly, and the boards are small enough to run several back to back. Different game, same three habits.