I also misunderstood this one at first. The mistake is assuming what “tolerance” is.
Look at this line:
Look at this instruction:
For example, if the tolerance is 0.01, the bisection method will keep halving the interval until the difference between the upper and lower bounds is less than or equal to 0.01.
In other words, keep halving WHILE the difference between the upper and lower bounds (high and low) is greater than the tolerance.
Does it make more sense?
This instruction I think is misleading maybe it should be removed
The tolerance being the acceptable error margin for the result.
An “error margin” seems to imply the result will be correct within plus or minus the tolerance.