Exercise 3: Why Keep Newton's Method as the Primary Solver, With Bisection Only as a Fallback? — Possible Solution ==================================================================== WHAT WOULD BE LOST BY REPLACING NEWTON'S METHOD OUTRIGHT ------------------------------ Chapter 7 verified a dramatic speed difference on the identical equation and tolerance: bisection needed 39 iterations while Newton's method needed only 4. For the large majority of valve configurations - the ones where Newton's method behaves normally, which based on this chapter's own framing is presumably most of them, since the cycling failure was specifically called out as affecting "one particular valve's calibration curve" - replacing Newton's method entirely with bisection would make every one of those ordinary cases roughly ten times slower for no benefit, since they were never at risk of the cycling failure in the first place. WHY A FALLBACK CAPTURES THE BEST OF BOTH ------------------------------ The set-point solver's actual problem was not that Newton's method is unreliable in general - it's that Newton's method has no built-in safeguard against the specific, verified failure mode (a permanent cycle) that this chapter's Step 6 diagnosis identified for one known input. An iteration cap with a bisection fallback preserves Newton's speed advantage for the common case (it's tried first, and normally succeeds quickly) while adding exactly the guarantee bisection offers (per Chapter 7's own explanation, rooted in the Intermediate Value Theorem) for the rare cases where Newton's method fails to make progress within a reasonable number of steps. WHY THIS IS BETTER THAN "JUST USE BISECTION EVERYWHERE" ------------------------------ Always using bisection would trade away Newton's method's genuine, verified advantage across the entire fleet of valves just to guard against a failure mode that Chapter 7 showed only affects specific, identifiable situations (functions with double roots, or starting points that happen to land in a cycle). The combined approach only pays bisection's slower cost on the rare occasions it's actually needed, rather than on every single set-point calculation the system ever performs. WHY THIS WORKS AS AN ANSWER ------------------------------ The explanation quantifies the real cost of abandoning Newton's method (the verified 39-vs-4 iteration gap) rather than treating "just use the safer method" as costless, and explains specifically why a fallback design captures Newton's speed for the common case while still eliminating the specific verified failure mode from Step 6, rather than trading one risk for a different, more pervasive cost.