🎲 Play Whe Dashboard

Top 12 numbers for a draw window and day of week — the desktop "By Draw Time" and "Longest Gap" tabs in one. Switch the mode to rank by current drought instead of hits.

What tends to come in the very next draw after a number plays (year-to-date (2026), all windows).

Rolling hit rate over the previous 60 draws (all windows) across the year — the dashed line is the 1-in-36 expectation. Flat wiggles around the dashed line = randomness doing its thing.

Hits per window, year-to-date (2026). Darker = more hits.

Folk family groupings ranked by hits per member.

Do the draws deviate from a fair lottery? Formal chi-square tests, FDR-corrected (running several tests at once guarantees a small raw p somewhere) — ⚠ only fires below adjusted p 0.05.

🎯 Bias monitor: No exploitable bias detected. A bet only turns profitable if some number hits above 3.85% (fair: 2.78%). Closest posteriors: #34 62% · #05 41% · #16 41% — but the luckiest of 36 FAIR numbers shows ≈56% at this sample size, so that's selection, not signal. Evidence so far: 765 of ≈3,786 draws (20%) needed to detect a break-even bias — this answer sharpens automatically as draws accrue.

Same formula as the desktop app (smoothed, optional exponential-decay recency), recomputed live in your browser. A heuristic lens — draws are random, this is not an edge.

Walk-forward: for each of the last N draws, rebuild the top-12 from prior draws ONLY (using the slider settings above, all windows) and check whether the actual draw landed in it. Random baseline 12/36 = 33.3%.

Tap numbers, set a bet per number, and see what you'd have won against the latest draw on file (33 Spider, 2026-08-18 Evening).

Every Play Whe bet has the same expected loss per dollar; what changes is the SHAPE — few numbers at big stakes swing hard, many numbers at small stakes grind smoothly. Compare at equal spend (assumes stake = spend ÷ numbers).

Prediction can't beat this game, but a mispriced PAYOUT can: break-even is $36 back per $1 (a 1-in-36 shot). If a promotion's effective payout — bonus included — crosses that line, the bet genuinely flips to +EV. Standard $26 never does.

Log plays for upcoming draws; they grade automatically against the draws on this page (revisit after the site updates). Stored in your browser's localStorage — clearing site data erases it.

Each position's digits blended hot + due (balanced, smoothed — the desktop app's formula). Top-3 digits per position shown as chips. A heuristic lens, not an edge.

How often each digit 0–9 lands in each of the four positions, year-to-date (2026). Darker = more hits.

Digit ranking for a position, draw window, and day — the desktop Pick 4 "By Draw Time" tab.

Co-occurrence of the digit in position A (rows) with the digit in position B (columns). Lift = observed ÷ expected if the positions were independent — blue above 1.0, red below, near-surface around 1.0. Raw shows plain counts.

Pick four digits, choose ORDER (exact) or ANY ORDER (box), and see what you'd have won against the latest draw ().

Expands the top digits per position — from the Best Bet consensus, matching the desktop Wheel tab — into playable lines. ANY ORDER merges duplicate digit sets; * flags sums outside the common band.