What the feed is made of

Three ways a feed is assembled

The difference between them matters to a reader because it decides whether the rows are a sample or a selection.

1. A curated list. A page that is written by a person or a rule, updated on a schedule, and typically keeps a row or two drifting in from earlier windows. This is a publication, and its rows are chosen.
2. A filtered stream. Rows pushed automatically from the settlement log as they clear a threshold. This is mechanical, but the threshold is still a rule of the display rather than a property of the casino.
3. A live tap on the log. Every settle above the bar appears as it happens. This is the least edited form, and it still shows only wins above the bar - which is the whole point, and the whole gap.

What one row records

A row on the sample feed carries four fields: a stake, a return, a multiple and a time. Two of those are derived, and none of them is a rate.

The four fields of a row, and the two that are missingshown    stake        1.00
shown    return     250.00
derived  multiple   250x        (250.00 / 1.00)
derived  time       14:07
missing  attempts   the spins behind this one row
missing  population the players those spins were spread across

The multiple is the return divided by the stake - a row can be re-derived from two of its own fields, which is why a feed feels checkable. What cannot be derived from a row is how many spins it took, and that is the number that turns a fact into a rate.

The feed against the hour it was drawn from

The sample feed holds twelve rows. In the same hour the casino settled 457 wins that cleared its own 250-times bar, across 4,750 players and 1,900,000 spins. The display is therefore a 2.6% sample - and it is not a random one, which the next pages measure.

Where to go next