d[e, 1, t] - d = a[-f, 1, t]
(-f) : n <-(e) : n+d
(-f) : n <-(e) : n+a+x
(e-(2x+1)) : n+a <-(e) : n+a+x
(e-(2n+1)) : d <-(e) : n+d
(e-(2a+1)) : n+x <-(e) : n+a+x
There are lots of other superpositions that you don’t have to calculate to use the existence of, since they all stem from e. Construct a root finding algorithm like the square root one. You have been given a 0 estimate.
Progress like so:
{e:n0 (from -f):d0:x0:a0:b0} -{e:n1:d1:x1:a1:b1} -> … -> {e:n:d:x:a:b}
This is the exact same thing as the n0 triangle bases concept and the intercept course concept. They are different ways of doing the same thing, which is progressing toward the correct answer by starting with a decent estimate and calculating error values to progress (in a complexity that is like binary search) to the right answer. Imagine the movements like one of those casino machines picking each value to the right answer one by one.
That link those concepts for you?
P.S. the p of t can be used as a way to calculate an error value (amount to change the estimates by)