Spunk Odds Reference

Scratch cards guide

Eagle Scratch: what a simulation of many sessions shows

Eagle is the name on this page's file, not a ticket brand or theme. The page runs a seeded simulation of scratch-card buying and explains how to read what comes out of it.

Why simulate at all

Formulas give averages and spreads. A simulation gives something more concrete: a set of plausible sessions, each with its own ups and downs. It cannot replace the exact maths, but it makes the shape of the results visible. The trick is to read it as an illustration, not as evidence about any real buyer.

The input

The simulated tickets are drawn from this invented run of 200,000 cards at 2 units. Each simulated ticket is a random pick from the full run, so the sessions behave like independent draws.

Hypothetical scratch-card print run of 200,000 tickets at 2.00 each
PrizeTicketsOddsExpected value per ticket
4,000.0031 in 66,666.70.0600
400.00301 in 6,666.70.0600
40.005001 in 400.00.1000
10.004,0001 in 50.00.2000
4.0012,0001 in 16.70.2400
2.0030,0001 in 6.70.3000

46,533 winning tickets out of 200,000 (overall 1 in 4.30). Prize fund 192,000.00 on sales of 400,000.00: return 48.00%, so a ticket at 2.00 is worth 0.9600 on average. Hypothetical print run.

The output

Below are eight simulated sessions of 60 tickets each, generated with a fixed seed so the page always shows the same numbers. Results are in units of one ticket price, so a net of -10 means the session lost the price of ten tickets.

Simulation: scratch, 8 x 60 rounds
SessionRounds wonNet (units)Lowest pointLongest losing run
19-45.00-46.0015
215-43.00-43.0012
314-38.00-38.0011
414-17.00-25.009
513-45.00-45.0016
611-25.00-25.0011
713-38.00-38.0011
813-43.00-43.0012

Seeded pseudo-random simulation (Python random.Random, seed 7731), 8 sessions of 60 one-unit rounds. Average session result -36.75 units. This is a simulation, not a record of real play; a different seed gives different numbers.

How to read the columns

Expect sessions to cluster around the expected loss, with an occasional outlier when a 40-unit or larger prize lands. The average of a handful of simulated sessions will not match the exact figure; with only eight sessions, one lucky ticket shifts it a lot. Change the seed and the table changes, while the exact expected loss of 52.00% of the price per ticket stays put.

What a simulation cannot do

It cannot prove the expected value, only illustrate it. It cannot predict a session, since a real one has no seed you can read. And with a top prize at 1 in 66,667, a small simulation will usually show no top prize at all, which is itself an accurate picture of what most buyers experience.

Glossary

Seed
A starting value for a pseudo-random generator. The same seed always reproduces the same simulated sessions.
Print run
The full, finite batch of paper tickets for one game, with a fixed number of prizes in each tier.
Return to player (RTP)
Total prizes divided by total sales. In this run it is 48.00%.
Break-even prize
A prize equal to the ticket price. It counts as a win but leaves the buyer where they started.
Session
A fixed number of tickets bought one after another; 60 in the table above.