There are subtleties, however. When the demand is varying it may be more economic to keep two generators on half load than to run one on full load and shut down the other which may then have to be re-started later. The capacity of the transmission system must also be taken into account, which may mean plant is "constrained off," simply because the grid is operating at maximum capacity in a particular region.
With a "free" market operating in the UK, the procedure differs in two key respects. First, all generators are paid the "system marginal price" and not their "bid" price. Second, capacity payments are built in to enable generators to recover their fixed costs. Elsewhere these are settled by annual charges. Fixed costs are calculated by multiplying the loss of load probability by the "value of lost load." However, for this to work properly, the plant margin needs to be optimal at all times -- a sticky question discussed in the main story.
Both these payment procedures have caused problems. Improvements to the operation of the pool are now being discussed with a weather eye on price variations. Fig 2 gives these for a typical recent winter day where the peak, reached at about 18.30, reflects the operation of the capacity credit mechanism.