Computer Week · 08 · A finite operating world

Bowling world

40 LANES · ONE OPEN DAYWORK LEDGER IN MOTION

In useFull rackReady / inactive

Edit the model assumptions

Illustrative inputs · values per open day unless labelled otherwise

A rack is not a delivery

Begin with a full rack ready and finish ready for the next game. After each delivery count either a full-rack event or a standing-pin service. This is an accounting convention; actual machines have short cycles, detection strokes, ball returns and motor timeouts.

History	Balls	Full racks
Twelve strikes	12	12
Ten open frames	20	10
Nine strikes; 0, 0	11	10

A bounded choice

The search varies three skill profiles, five active-lane counts, two machine types, three oil intervals and three release cadences: 270 policies. Edits to strike probability, spare probability or training cost affect only the currently selected skill's 90 policies. The other two profiles retain their defaults. Choosing another skill or restoring its preset clears those profile edits.

All policies retain the current demand, prices, fixed costs, ball speed and other assumptions. Explicit edits to machine service prices, cycle times or electrical use apply to both hardware choices. Unedited coefficients retain their hardware defaults.

The green connecting line is the cost-versus-games Pareto frontier among policies meeting the service floor. A point belongs there if no feasible policy serves at least as many games at no higher cost, with one strict improvement. A selected objective chooses within this grid, not among every possible business.

Click a point or use the arrows to inspect. Applying it changes the operating controls. A policy can be cheap by leaving demand unserved; the service floor makes that choice explicit.

Replace these numbers

Urban and rural are scenario names, not empirical estimates. Actual booking records, ball and cycle counters, downtime, labor schedules, invoices and condition measurements would be needed to calibrate this model.

Strike and spare probabilities run from zero to one; the spare probability is conditional on a nonstrike first delivery. These two inputs and training expense edit the selected skill profile, including its policies in the search. Switching skill presets clears those edits. Machine cost, cycle-time and electrical edits apply to both hardware choices in the search.

Choose demand in player-games or occupied lane-hours independently of billing per game or per lane-hour. Faster service does not create sales beyond that demand. A minimum service fraction constrains the policy search.

Every legal pin-count history