Numbers method (Study tier)

In short: this section does not predict which numbers “must” appear. It shows which numbers lag or lead archive statistics, and helps you build a set for a ticket or the basket.

1. What the method does

Each of the 49 numbers gets one of three statuses:

From that you get a ready set you can view on layouts, check with a backtest, and send to the basket.

2. Setup (Study)

1. On the left, pick a draw snapshot and archive year (“Archive from year”). Frequencies and methods use the archive from that year.
2. Open the Numbers section.
3. Choose backtest depth: 10 / 20 / 30 draws. Study unlocks all three; the on-screen rating is scaled to a “per 10” baseline.

3. Settings mode

Standard

• Fixed: 16 recommended / 16 not recommended / period step 3 months.
• At the top — a best method card from the precalc for the selected year and depth.

Advanced — unlocked on Study

• Sliders: how many recommended (1–35), how many not recommended (1–24), period step 3 / 4 / 5 / 6 months.
• “Use best parameters” — loads the best OR settings from the precalc.
• Full year-rank tables are available.

4. Six ways to build a set

A. Archive frequency (min / max)

• Green: the rarest in the archive.
• Red: the most frequent.
• Play sets are usually taken from the greens.

B. Cumulative period (from January)

Compares against yearly averages from the selected archive year. Watches pace from January in 3 to 6 month steps.

• Green: by period end they have not yet reached the quota.
• Red: already ahead of the quota (with margin).

C. All recommended (intersection “AND”)

Green only if a number is both rare in the archive and lagging the period.

Used on large sets. The set is smaller and more strictly filtered.

D. All not recommended (intersection “AND”)

Red only if both frequent in the archive and ahead of the period.

Used when you deliberately look at the hottest numbers’ hits.

E. All recommended (union “OR”) — often best in the ranking

Green: rare or lagging the period.

Wider set; in Advanced the system often suggests this one (prediction efficiency up to 150% — 1.5× better than random selection).

F. All not recommended (union “OR”)

Red: frequent or ahead of the period.

Used when you deliberately look at “warm and hot” numbers’ hits.

Study practice:

1) “Best method” card → 2) “Use best parameters” if needed → 3) confirm the set on a layout → 4) to basket.

5. How to view the set (4 layouts)

1. Ticket — grid like a Bonoloto slip.
2. Circle — recommended in the centre, not recommended outside.
3. Trapezoid — recommended on top, not recommended below.
4. Two columns — recommended left, not recommended right.

Number tooltip: how often it appeared and how long ago (on the Ticket tab); in the period method — also the period quota/norm.

6. Where the set goes

1. Under the layout: “Add to basket” — the full current recommended set (or not-recommended if an avoid method is selected).

Study practice: with Study membership, in the basket you can combine pools, scenarios, intersections, Auto Pick, and more.

7. Checking the method (backtest)

On the last draw

Button “Check method on the last draw”.

The method is computed without the last draw and compared to it.

Batch backtest

• depth 1 — visual check only;
• depth 10 — shallow assessment;
• depth 20 — still a small randomness factor in the method;
• depth 30 — practically removes the randomness factor of the method you chose.

Then:

1. Pick depth at the top (Study — up to 30).
2. Set the archive year for training.
3. “Run backtest”.
4. Read the table: date, forecast, hits, draw, 5+C and rating.

8. Typical Study workflow

1. Archive from year + depth 30.
2. Advanced mode.
3. Use best parameters (or set rec/avoid + step manually).
4. Most often — recommended “OR” (if the ranking confirms it).
5. View on Ticket / Circle.
6. Batch backtest on 30 draws.
7. If it looks good — to basket, optionally intersect with a pool/scenario in Compare or Auto Pick.

Important

• Every 6-number combination has the same odds; the method only orders archive statistics and groups statistically equal numbers into sets.
• “Recommended” ≠ “will hit tomorrow” — these are numbers that lag the chosen model.
• Method efficiency — the best set quality signal; checked via backtest results.
• Set rating — another integrated performance score (not tied to archive stats, only to backtest results).

Study gives the full loop: Advanced, depth 30, best parameters from the full archive/method grid, year rankings, and a convenient check before you play.