What is K-means clustering?

K-means is a machine learning algorithm that automatically groups objects into K clusters by similarity of characteristics. For lottery numbers, three characteristics are used: draw frequency, depth, and average interval.

What do the cluster names mean?

Names are assigned automatically based on average characteristics: "hot stable" — high frequency, low depth; "cold stable" — low frequency, high depth; "warming up" — low overall frequency but low depth (recent appearance).

How many clusters should I choose?

4 clusters is recommended for standard lotteries. For lotteries with many numbers (e.g., 6 out of 49), you can try 5–6 clusters for finer grouping.

Does cluster analysis guarantee a win?

No, cluster analysis is a statistical analysis tool, not a prediction tool. It helps structure information about numbers, but each lottery draw is an independent random event.

Lucky-numbers.ru

Lotteries Randomizer Premium Access

New Zealand: Strike

Strike Cluster Analysis — Grouping Numbers by Behavior

Strike: k-means clustering: combining numbers with similar draw patterns

Cluster analysis groups Strike lottery numbers by a combination of characteristics: draw frequency, depth (recency of last draw), and average interval. Instead of analyzing individual metrics, you see the complete picture: "hot stable", "cold", "warming up", and other groups.

Analysis based on 20 draws from 03/18/2026, 07:20 AM to 05/23/2026, 08:00 AM

Clusters:

Hot stable

Numbers: 9

Average frequency: 4

Average depth: 1.8

238121422242834

Warming up

Numbers: 10

Average frequency: 1

Average depth: 4.1

4101115161820212632

Cold stable

Numbers: 7

Average frequency: 0

Average depth: 18.4

13232735363739

Medium

Numbers: 14

Average frequency: 2

Average depth: 8.6

15679171925293031333840

Cluster Visualization

Frequency vs Depth — color = cluster

Clustering Table

All numbers with their cluster and characteristics

Added to generator 0 / 40

Selected 0

Ball added	Cluster	Frequency	Depth
Add	Hot stable	3	3
Add	Hot stable	3	4
Add	Hot stable	3	1
Add	Hot stable	4	4
Add	Hot stable	5	1
Add	Hot stable	4	0
Add	Hot stable	3	0
Add	Hot stable	3	3
Add	Hot stable	5	0
Add	Warming up	1	5
Add	Warming up	1	1
Add	Warming up	2	0
Add	Warming up	1	6
Add	Warming up	1	5
Add	Warming up	2	4
Add	Warming up	1	1
Add	Warming up	1	7
Add	Warming up	1	4
Add	Warming up	1	8
Add	Cold stable	1	17
Add	Cold stable	0	20
Add	Cold stable	0	20
Add	Cold stable	1	13
Add	Cold stable	1	19
Add	Cold stable	0	20
Add	Cold stable	0	20
Add	Medium	3	8
Add	Medium	2	11
Add	Medium	2	2
Add	Medium	2	7
Add	Medium	2	10
Add	Medium	2	2
Add	Medium	2	9
Add	Medium	3	6
Add	Medium	3	11
Add	Medium	2	14
Add	Medium	2	7
Add	Medium	2	11
Add	Medium	3	12
Add	Medium	2	11

Generator by Cluster Analysis

Mark numbers from clusters of interest

Selected numbers for generator

G on keyboard — generate combination

No combinations

Generate, load or paste combinations.

How to Use Cluster Analysis for Strike

Step-by-step guide to clustering lottery numbers for Strike lottery

Choose the field and number of clusters

For multi-drum lotteries, select the desired field. The number of clusters (3–6) determines grouping detail — start with 4.

Study the cluster cards

Each cluster has a name describing the numbers' behavior: "hot stable", "cold stable", "warming up". Pay attention to the average frequency and depth.

Analyze the scatter plot

The "Frequency vs Depth" chart shows how numbers are distributed among clusters. Numbers in the top-left are frequent and recent; bottom-right are rare and long absent.

Use the generator

Mark numbers from clusters of interest and generate combinations. For example, choose "hot stable" for an aggressive strategy or "warming up" to bet on the trend.

About Cluster Analysis

K-means method and its application

K-means is one of the most popular unsupervised machine learning algorithms. It groups objects (numbers) into K clusters by minimizing the distance to the cluster center.

Frequency

How many times the number has been drawn

Depth

How many draws ago was the last appearance

Interval

Average gap between appearances

Comprehensive approach

Cluster analysis evaluates a number not by a single indicator, but by a combination of characteristics. A number with high frequency but great depth is not the same as a number with high frequency and low depth.

Cluster Analysis Tips

Practical recommendations for effective use of the method

Start with 4 clusters — this is the optimal number for most lotteries. Increase only if you need more detailed grouping.

Pay attention to the "warming up" cluster — these are numbers that have recently started appearing more frequently after a long quiet period.

Combine numbers from different clusters to create a balanced combination.

Clustering results depend on the number of draws. Use the period filter to compare results across different time intervals.

Frequently Asked Questions about Strike

Answers to common questions about lottery cluster analysis for Strike