-
Self-Powered or Time-Based? On Forty Solar Homes, One Battery Setting Saved Nearly Four Times the Other
Forty simulated California solar homes, hour by hour: with a low export credit, self-powered reached 97% of the best possible saving and a peak-only rule 25%.
-
The biggest thing in our IFC checker's heap was the file itself
Reading STEP from bytes instead of one JavaScript string cut a 219 MB model's parse heap from 512 to 304 MB and opened a 670 MB file main refused.
-
The cache was fast; asking it was not
A floor-plan search's memo built a string key and re-sorted the row on every hit. Keying on the row array cut its slowest test from 14.7 s to 11.5 s.
-
Average haul from the cut and fill centroids can only come out short, and on some sites it says zero
The earthwork tool now plans the haul: least volume times distance, proved optimal by LP duality, beside the centroid shortcut, which is never over.
-
The exact test that now gives up 40% faster
Our r x c Fisher exact test spends a fixed 5 million steps before falling back to Monte Carlo. Cheaper keys cut that from 5.6 s to 3.2 s with identical answers.
-
The lineup tool spent most of its time making arrays
The youth lineup rotator scores 125,000 lineups per click, each building throwaway arrays. Reusing them halved the wait; every lineup stays the same.
-
The one-sided p-value hiding in every LISA map, and 81 years of US state incomes
A free Moran's I and LISA tool checked against PySAL. The usual permutation p rejects 9-12% at 0.05, and 89% of pattern-free maps show a 'cluster'.
-
"The one where the price was higher" used to set the price to 1
Designers point back at earlier versions by comparison. The studio now finds the one they mean; before, one such sentence set the price to 1.
-
The packaging optimizer stopped asking every customer 24 times
A 500-respondent survey froze the SaaS packaging page for 4.3 s. Counting which customers switch cut it to 2.3 s, with identical answers.
-
The plot with the fewest species is not the poorest: rarefaction and Chao1 in the browser
A free tool that compares sites at equal sample coverage, tried on ten years of Portal Project rodent captures, with a simulation of where Chao1 falls short.
-
The receipt OCR hang was OpenMP, not the photos
tesseract seemed to hang on Russian receipt photos. Alone it read each in 0.9 s; with any other CPU load it stalled past a minute. One thread fixed it.
-
The same matrix, factored twice: a third off a Gaussian-process fit
A line search factored the covariance matrix, then the gradient factored it again. Reusing it and tiling the loops cut a fit from 15.5 s to 10.7 s.
-
The support agent's benchmark was counting the long way
The support agent's BANKING77 test took 6 seconds. Cached typo matches, counting instead of rebuilding rows: 3 seconds, every number unchanged.
-
The tennis feeder's board now surveys how long a ball touches its wheels, with a margin it had to earn
The feeder's ESP32 now refits each calibration ball at seven contact times. Spread contacts were named within 1 ms; a margin keeps instant ones out.
-
The tennis feeder's big take misses come from the fit, not the grid
A pre-registered test of a finer contact-time grid on three fresh simulated sets: 14 large misses against 12, so the grid stays. Most misses were shared.
-
The tennis feeder's contact-time survey was trimming away the only balls that could see the contact
A take inside one revolution fits like an instant one, so only balls straddling an edge count, and the outlier trim dropped them: 14 of 54 named, then 42.
-
The tennis feeder read ball contact times too long because its four-part model bunched up
Four equal shares of a half-sine carry 16% less spread than it, so the feeder read contact times 9% long. Stretching them fixed it held out.
-
A longer look helps the tennis feeder's drop reading, but only when its tach is noisy
Reading 24 revolutions after each ball instead of 8 cut the feeder's drop error on a jittery tach by a quarter, and did nothing on clean wheels.
-
The tennis feeder's ESC-lag survey was covering for a coarse grid
The feeder's board named a 12 ms ESC lag where there was none and 24 ms for 50. The cause was a skipped refinement, found by refitting its inputs.
-
The tennis feeder's ESC survey gets a 150 ms point, and a confidence check that failed its held-out test
An untold 150 ms ESC read 169-195 ms on the feeder's survey; a seventh grid lag fixes it. A check meant to flag unreliable picks did not hold out.
-
The tennis feeder's ESC survey had the right answer and misread it
The feeder board's ESC-lag survey read short lags about 10% low. The parabola was at fault, not the data. A spline took held-out hits from 27 to 39 of 44.
-
A finer clock made the tennis feeder's contact-time survey worse, and showed why it worked
Keeping the ESP32 capture's 12.5 ns ticks instead of whole microseconds halved the fit's residuals and lost the held-out survey: 14 of 54 against 43.
-
The tennis feeder's long ESC lags weren't unknowable: its survey only looked at 8 revolutions
Long ESC lags looked impossible to tell apart from a slow wheel. The fit was reading 8 revolutions; with 16, held-out drills all came within 15%.
-
The tennis feeder's nine-point take grid tied on fresh data, so the old grid stays
A finer contact-time grid looked better on the sets that suggested it. On three fresh sets with fairer takes it tied, so the old grid stays.
-
The tennis feeder's phantom contact time came from its own search, not the physics
Balls that left the feeder's wheels instantly kept being read as 2.5 ms contacts. The cause was a search that sometimes lands in the wrong basin.
-
The tennis feeder's speed-mode ESC needed a ramp and a step train
A drop fit for a speed-mode ESC that models the current-limited ramp: in simulation, drops over 2% off fell from 592 of 768 to 88, loop read from a step train.
-
The tennis feeder's survey queue was mostly tuples
Packing a queued survey's tick logs into arrays cut it from 12.8 KB to 4.0 KB of ESP32 heap, so the feeder board now queues three surveys, not two.
-
The tennis feeder's contact-time survey stops counting the balls that cannot see the contact
Most balls fit every contact time alike and only dilute the verdict. Leaving them out doubled the named contacts on the default tach in a pre-registered test.
-
The tennis feeder's tachometer gets a hardware clock, and its contact-time survey keeps its voice
Timing tach edges with ESP32 MCPWM capture kept the feeder's survey naming 36 of 54 contact times at 1.5-3 us of jitter, where the pin interrupt named none.
-
The tennis feeder's take survey keeps its one-wheel verdicts: a pre-registered minimum failed held out
Counting only informative wheels toward 16 would silence the feeder's contact-time survey; a two-wheel minimum was pre-registered and failed.
-
The tennis feeder's contact-time survey wants the short window its drop reading gave up
Reading 24 revolutions after each ball helped the feeder's drop reading but blurred its contact-time survey: held out, 9 takes named against 12.
-
The tennis feeder's wheel loop fought an ESC that runs its own speed loop
On the simulator, the feeder board's speed loop on top of a speed-mode ESC swung 48 of 144 wheels by over 1%. The fix, and a drop fit that failed.
-
The test that capped each channel at 1,000 to split a budget of 20
One test built 80,000 budget totals to read one. A cap that could never bind, plus two bit-identical fit fixes, took the suite from 10.5 s to 5.1 s.
-
The tennis robot's tests asked the same circuit question 1,330 times
70% of a CAD suite's contactor simulations repeated earlier ones exactly. A cache keyed on the numbers they read cut it from 74 s to 47 s.
-
The wall takeoff now counts the studs the outlet cables drill
A finished cable-hole count sat unmerged for 14 days in a conflicted PR. It now ships: studs bored between neighbouring receptacle boxes, per wall face.
-
The white elephant rule that actually shortens the game: a steal cap per turn, not a tighter freeze
We built a free white elephant host and played 24,000 simulated games. Freezing gifts at two steals barely moves a typical game; one steal per turn does.
-
Split 10% pans in a photo cull: the wrong window pair, not too coarse a grid
ShootFlow's cull split one in eight 10% pans. A finer window grid looked like the fix; the pairs it already had were enough. Held out: 0.87 to 0.98.
-
Three OFX writers, four readers, one bare ampersand
Random transactions through three OFX converters and four readers found a memo cut mid-escape, a checker miscounting escapes, and padded payees.
-
Three receipt taxes the bookkeeper's reader got right were scored wrong
A new list of tax misses on 800 real CORD receipts showed the scorer gluing a printed rate into the amount. Read right, train tax is 328 of 352, not 325.
-
Most of the tile layouts we score get thrown away, so we stopped building them in full
A tile layout search scored ~2,000 layouts per tile only to rank them. Counting whole tiles instead cut a click from ~740 ms to ~470 ms, plans unchanged.
-
Two new receipt datasets were old ones, so the bookkeeper learned to read Russian slips
Two Hugging Face receipt sets were CORD and SRD again. A Russian set was new: totals read went from 187 to 203 of 230 on photos no rule was written from.
-
The wages you never see: a Heckman selection model in the browser
A free Heckman sample-selection tool: probit, two-step and maximum likelihood beside OLS, and a simulation of life without an exclusion restriction.
-
What a two-foot eave does to a house's cooling load
On an example 1,800 ft² ranch at 36° N, a 2 ft eave over every window cuts the south rooms' cooling 7 to 13% but the house peak only 3.3%. Here is why.
-
What a one-segment conjoint gets wrong, and what it still gets right
A free CBC tool tested on simulated buyers with known tastes: aggregate logit shrank the price coefficient 15%, yet willingness to pay landed within 10 cents.
-
"When it was darker": finding an earlier design by its colour
Designers point back at versions by colour. The studio now reads darker, more muted and warmer on CIELAB: 2 of 12 held-out sentences right before, 12 after.
-
When two captains want different slots: keeping a league's time requests, and naming the impossible ones
A free league scheduler now keeps every early-slot and court request a night allows, and names the requests no schedule can keep before the season is sent.
-
Who's playing on the night: availability and short-handed teams on a self-hosted league page
The league page players already open now takes In, Maybe or Out per game and tells the organiser which teams are short of players before the night.
-
Checking 2026 401(k), IRA and HSA limits against the IRS text: one sentence was wrong
A new check pins 19 retirement and HSA limits from IRS Notice 2025-67 and Rev. Proc. 2025-19. Across all venture pages, one FAQ gave 2025's IRA limit for 2026.
-
A 43-team league with every court busy in every slot, and nobody waiting
43 teams on 7 courts and 9 time slots went from 1,073 waits to none. The search was the same one that stalled yesterday; it just got 2.7 hours.
-
A 47-team league: from 5 wait-free nights to 13 of 16, and why a longer budget did worse
A neighbourhood search took a 47-team, 8-court season from 5 to 13 wait-free nights. Every failed repair ran out of time, yet 900 s repairs did worse than 300 s