Read your analysis in three steps

The latest official draw loads by itself in the white box. The button sits right under it. Your analytical outputs appear in a few seconds, built from the full EuroMillions record up to that draw.

Under the button, your run is laid out under five labels. These are the main outputs of every run. Each list sits in zone rows, 1–9, 10–19, 20–29, 30–39 and 40–50, and a number always keeps the same place, so you can compare the lists at a glance. Every number is shown as the same blue ball: a list tells you which numbers it holds, and none of them is marked out by colour.

• Historical Memory AI PriorAI candidate pool 25/50

  A trained memory layer reads the complete record and narrows the fifty numbers to a ranked pool of 25. The three models use it as their guide.

• A1 — Multi-Signal Deterministic ModelAnalytical set 12

  The balanced analytical set. It weighs how often and how recently each number has been drawn.

• A2 — Deterministic ModelAnalytical set 12

  The constrained analytical set. It reads which numbers appear together and follow each other, with a tighter spread.

• A3 — Zone/Split Markov learnerAnalytical set 12

  The exploratory analytical set. It reads how numbers move between the zones and split zones of the board from draw to draw.

• Unique numbers (35)

  The AI candidate pool, plus the numbers A1, A2 and A3 add outside it, taken in that order up to 35 numbers in total. It is the widest view on the page.

A simple way to read any run.

1. Start with the AI candidate pool

   The Historical Memory AI Prior is the field the three models work from. In this run it holds 25 of the fifty numbers.

2. Compare the three readings

   A1 is the balanced reading, A2 the constrained one and A3 the exploratory one. Tap a number to see which of the lists hold it. Where they differ, the descriptions above explain why.

3. Check the spread

   The zone rows show how the numbers of each list sit across the board, from 1–9 to 40–50.

4. Use Unique numbers for the widest view

   It brings the four lists above it together, each number once, up to 35 numbers. In this run it holds 35.

Under the outputs, the workspace brings every model output and its record together. It works in four steps.

Step 1. Review a model output and its record. Each card is one model output: Historical Memory AI Prior, A1, A2, A3 and Unique numbers. The large figure is the size of its current draw analysis; the line beneath it is its record across your scored runs. Select a card to open that record. Every saved run is scored against the official draw that followed it: the drawn numbers the output held turn blue, the rest stay empty, and the score sits on the right. A run still awaiting its draw is marked with a gold pulse and is scored automatically once the draw is published. Until your first run is scored, the record shown is the public Tychefy record from the History page.

Step 2. Select 5 numbers for your analysis. White balls are the draw analysis of the output you selected; empty balls lie outside it. Tap a ball and it turns blue; once five are selected, the remaining balls dim. Your selections collect in the bar beneath the board, and Model agreement shows, for each of them, how many of the four models (Historical Memory AI Prior, A1, A2 and A3) include it and which outputs carry it.

Step 3. Select 2 Lucky Stars. The Lucky Stars follow the same method, with their own five outputs and record: AI candidate pool, Standard Top-2, A. Zone-diversified Top-2, B. Direction-diversified Top-2 and C. Mu-band Top-2. White stars are the draw analysis of the output you selected; tap a star and it turns gold. Model agreement shows how many of the five outputs include each Lucky Star you select.

Step 4. Build your analysis set. Add the selection to your analysis set, where your numbers appear in blue and your Lucky Stars in gold. Copy the set when you are done.