Posted in未分类
Why 8 Clean Outs Beat 31 Dirty Outs (Poker Equity vs Rule of 4 and 2)
The Rule of 4 and 2 is right — it just hides the dirty-outs problem. 17 clean outs vs 6 dirty on Ah Kh / Qh Jh 2c, with 60% equity vs the textbook 36%.
Posted in未分类
Why 1 kHz Is 0.343 m in Air and 5.96 m in Steel: The Frequency–Wavelength Trap
A 1 kHz signal is not the same length in every medium. The wave equation v = f · λ plus four medium presets (air, water, steel, light) decide everything from antenna length to sonar resolution — and a free browser converter handles the math in under a second.
Posted in未分类
Three Regexes Catch 95% of PDF PII — And the Two That Always Slip Through
The default pattern set leaves your team compliant on the day you enable it and exposed the day you forget a field. Here is the two-pass review workflow that survives a real audit.
Posted in未分类
Where Your Contrast Checks Stop at 4.5:1 and the Heatmap Keeps Going
A spatial WCAG audit replaces per-pair ratio checks with one overlaid view of every failing region — including the gradient failures no eyeball test catches.
Posted in未分类
AI RAG Chunk Quality Scorer: The Four Metrics That Decide Whether Retrieval Answers Anything
Chunk size is the upstream lever that decides whether RAG retrieval answers the question. Score three candidate schemes on coherence, coverage, overlap health, and chunk-size consistency before you index.
Posted in未分类
Inside the HTTP/3 QUIC Handshake: 1-RTT, 0-RTT, and Reading a PCAP
A byte-level walkthrough of the HTTP/3 over QUIC handshake: 1-RTT fresh, 0-RTT resumption, connection migration, and how to read a packet capture without going cross-eyed.
Posted in未分类
API Latency Budget Planner: Why Most Budgets Are Fiction Until You Compare Them to a Trace
A 300 ms P99 budget that ignores OpenTelemetry measurements is a guess with a flag on it. The three decisions every team makes — cold path, warm path, processing — are almost always the wrong ones until the waterfall meets the trace.
Posted in未分类
Cron Expression Generator: The Five-Field Decision Nobody Reads on Monday
Six to nine characters decide whether your cron job fires every minute, every workday, or never. Here's the syntax that drives every Unix scheduler, and why a generator matters more than typing speed.
Posted in未分类
One Total, Four Scores: DOTS, Wilks, IPF GL and Goodlift Compared
An 83 kg raw lifter hitting 600 kg posts 405 DOTS, 400 Wilks 1994, 88 IPF GL and 88 Goodlift — three different tiers from one total. Here is why the four formulas disagree, what each polynomial actually does, and when to quote which score in programming, meet prep and gym-floor conversations.
Posted in未分类
Blue Hour Calculator: Why Photographers Live by a 20-Minute Window
Predict the precise blue hour window for any location and date — solar elevation math, IANA time zone handling, and a confidence indicator for polar edge cases.