You've added a variable that would be beneficial - that is, the supply voltage, or Vref. It is likely that to get the optimum tuning that you are in the process of determining, to get the upper end TPS voltage, you would need a negative resistance added between the top of the potentiometer and Vref. Since we know there is no such thing as a negative resistor, your jacking up of Vref is the perfect solution. From there, the real value resistors could be determined.
Fortunately I anticipated that when I built the formulas - I made the Vref as one of the variables, so all that would be required would be to pick a Vref voltage that is determined to be high enough, and go from there. That would allow you to raise the idle voltage and the WOT voltage at the same time using the variable resistors at both ends at settings determined from the formulas.
For your experimenting, it would probably be easier to set a fixed voltage for Vref determined to be high enough for a range of possibilities, and put in adjustable resistors on both ends - then dial in the bottom and top end TPS voltages by setting the variable resistors according to the formulas. That way you can quickly experiment with different settings of the two ends simultaneously across all operating conditions - not just trying to optimize bottom, top, and middle in separate test runs. Just a suggestion.
'98 LXi - Later Concorde gages (black w/ chrome rings)/'99 LX - LHS gages (white) - HIR bulbs