🔥❄️ Frequently Drawn / Cold Numbers (Hot / Cold)

This page splits digits into two groups: Hot (appearing unusually often) and Cold (gone unusually long) — using real statistics, not raw counts.

Hot is not just "appears a lot" — a digit must be far enough above expected that it is unlikely to be chance. We use a z-score (distance from the mean, measured in standard deviations).

Where 1.96 comes from: it is the industry-standard cutoff for "95% confidence" in statistics. In a normal distribution, 95% of values fall within ±1.96 standard deviations of the mean — so z ≥ 1.96 means there is less than 5% chance the frequency happened by accident (α = 0.05).

This same threshold is used in scientific papers, drug trials, and academic research — it is the "academically acceptable" line.

In the table, e.g. "appeared 25 times (expected 18)" means:

• expected = the count this digit "should" have if all 0-9 were equally likely · formula: draws × positions ÷ 10 · e.g. 30 draws × First Prize (6 positions) ÷ 10 = 18
• Actual 25 > expected 18 → 7 over → if the excess is large enough (z ≥ 1.96), it qualifies as Hot

Cold means a digit has been missing longer than usual. We compare draws since its last appearance against its theoretical average gap.

Overdue threshold: gap > 2 × average → only then does it flag "⚠️ overdue"

Where "avg" comes from: not from raw data, but from a probability formula: 1 ÷ (1 − 0.9N) where N = number of positions.

Real example (Last 2 digits): "gone for 21 draws (avg 5.3 draws), over average 3.99× ⚠️ overdue"

• 21 ÷ 5.3 = 3.99× ← how many times the average
• Overdue threshold = 2× → 3.99 > 2 → ✅ flagged as overdue
• Equivalent: threshold = 2 × 5.3 = 10.6 draws → 21 > 10.6 → overdue

Why 2×, not 1×? Because "average" is a midpoint — sometimes the gap is below average, sometimes above. That is normal. Going "more than 2× the average" is the band that is clearly unusual. · Note: overdue is NOT a guarantee that the digit will appear soon — each draw is independent.

In addition to the "aggregate" Hot/Cold cards above, there is a table that breaks the analysis down one position at a time — using the SAME data scope from the dropdowns above ("Prize Type" + "Time Range"). For example: First Prize + 30 draws → 6 rows (positions 1-6).

How to read it. Suppose row 1 shows:

Reads as: at position 1 of First Prize → the most-drawn digit is 7 (z = +2.34, clear standout) · the longest-gone digit is 3 (gone for 15 draws = 1.5× the average ~10 draws/position).

Two sides, two metrics (matching the aggregate cards above):

• 🔥 Hot uses z-score (count-based) — how much more often a digit appears than expected. Clear standout when z ≥ +1.96.
• ❄️ Cold uses overdue ratio (time-based) — how many times longer than average it has been missing. Overdue when ratio > 2× (avg per position = 10 draws).

Why it matters: some digits are hot only at specific positions — e.g. digit 3 might be hottest at position 5 but ordinary elsewhere → invisible in the aggregate view. This signal is useful when configuring "Custom Random" (Weighted), where you can pick a strategy per position.