"Wrong" is such a severe word. I tend to think of it as "conservative". Seems to me there are two schools of thought on this subject: is lift-off a primary stress issue (SUS) or is it secondary (EXP)?

For the pump example, stresses typically aren't the driving force in the design. Allowable pump nozzle loads generally dictate the use of spring supports or some other means to take the load off the nozzle. The multiple temperature example is different.

My approach: look at the SUS load that is being eliminated by lift off. If the "dead weight" load is small (compared to adjacent supports) I feel comfortable removing this from the analysis since the re-distributed load will be small. If the system is not overstressed without the support, this is conservative. However, if it fails, I place a spring support at this location, or find some means to re-distribute the loads so the system will not be overstressed.

A number of years ago, a client had an audit done on some compressor piping. The auditor used a load case combination that seemed very unconventional. The majority of the time I use the load case combination

1 W+P+T (OPE)
2 W+P (SUS)
3 L1-L2 (EXP)

or variations thereof. However, I believe the auditor used

1 W+P+T (OPE)
2 T (EXP)
3 L1-L2 (SUS)

or something like it.

Basically it boiled down to the issue at hand, dealing with lift-off. He was of the opinion that his load cases correctly dealt with lift-off. We agreed to disagree. Regardless, his comments were relatively minor and not related to overstressed points due to lift-off. We incorporated his comments to the satisfaction of the client.