It’s just reductionism: if changing data is bad, changing data is always bad.

Same nonsense gets you how terrible surgeons are (if cutting people with knives is bad, cutting people with knives is always bad).

It’s a good rhetorical technique because if you remove enough context you can pretty much always find something “exactly the same” to prove a point.

> It’s just reductionism: if changing data is bad, changing data is always bad.

It's not reductionism, and you're missing the point.

The whole point is that changing the original data to be able to support/reject hypothesis naturally raises the question of whether there is foul play.

If original measurements don't support your claim and suddenly by changing data you get it to fit your belief, any rational analysis would quickly flag the risk of data manipulation and scientific malpractice.

If instead you're dealing with a politically charged topic that attracts denialists and contrarians then you're making yourself vulnerable to accusations of fraud that will certainly be used to poison the well.

Do you understand why this is a problem?

> The whole point is that changing the original data to be able to support/reject hypothesis naturally raises the question of whether there is foul play.

People are applying a stronger point here where changing the original data is hard evidence of foul play and sufficient to completely discount any changes.

We'd be grappling with the reality of faster than light neutrinos now, though, if that was correct logic.

> People are applying a stronger point here where changing the original data is hard evidence of foul play and sufficient to completely discount any changes.

Manipulating field measurements is already frowned upon in all applications. There is no point. Things like data provenance is a serious issue, which is directly targeted by peer reviews and investigations on scientific malpractice. Even performance benchmarks highly favor standardized test and data sets.

Being objective matters. Once you start messing with original measurements, you place yourself in a position where you need to answer questions on whether you're just adapting data to fit your belief instead of the other way around.

In this case there is an observable relationship between wind speed and central surface pressure which has been observed over decades, but it has changed between the 1950s and the current day (although it has been stable across recent decades). The difference is in the measurement of wind speeds and with modern dropsondes we have much better measurement of wind speeds than existed in the 1950s. There is a clear and consistent bias in the windspeed-pressure relationship between decades. The correction that has been applied has been to apply a bias to correct the wind speed to the central pressure measurement. That correction was proposed in 2005 and so is nearly 20 years old at this point and had nothing to do with the current paper on cat 6 storms.

> The whole point is that changing the original data to be able to support/reject hypothesis

How do you know this? Do you understand the reasoning behind bias correction?

> Do you understand why this is a problem?

The problem is people do not understand bias correction and you are certainly not arguing on a methological level.