Predicting the Result of an Additional Trial (SIN 502)

A table of trials is just a record of what happened at certain settings. If those results move in a steady pattern, like doubling every 10 degrees or rising by a fixed amount each step, that pattern almost always holds for settings nobody tested yet. So you're not guessing. You're reading the rule the data already wrote and applying it one more time.

To predict a trial past the end of the table, find the step size, then count how many steps separate the new setting from the last one tested and apply the pattern that many times. If the new trial falls between two rows instead of past them, you don't need the step size at all. Just use the two surrounding values as a floor and ceiling: a steady trend can't jump outside its own neighbors.

The one twist sits at the far edge of the table. A line fitted to four or five trials only describes the range it was built from. Push the prediction far past that range and you're trusting the pattern to hold somewhere it was never checked, which is risky if the setup has a built-in ceiling or breaking point, like an enzyme that denatures past a certain temperature. This is where the skill connects to reading experimental limits elsewhere in the unit: a good prediction always asks not just what the trend says, but whether anything in the setup would stop the trend from continuing.