I used to be a big science fan

   Not possible.

   You conduct an experiment proving the twin paradox then.

   Oh, it is real enough.

   It happens often that when people "debate" religion they have something against religion that they believe "science" can solve. The so called "Creationism vs. Evolution" debate tends to pit science against religion as if religion is the problem everywhere because it always fails science. Few people have paid much attention to that particular debate lately, but the dynamic does tend to appear in the political arena.

   The problem with "science" is that when some people join the fray their arguments are based on an idealized "science" rather that how it plays out in reality.  They claim, "There is no faith in science, the focus is entirely on the results of real world testing."

   Obviously that is false. Of course there is an "ideal" that "science" always depends on real world testing, but much that passes for science in the real world in no way measures up to that ideal.

   It may be supposed that some people assume there is some "clean up crew" that catches all the errors "science" ever makes. Bountiful evidence however indicates that crew fails very much.

    The certainty of science ceteris paribus is rather often mistakenly attributed to statistical analysis. Medicine for example is far more statistical analysis than sceince ceteris paribus.

    The national debt and the high cost of health care indicate the voters are not exercising good science however much any party thinks it does. The reality is far short of the ideal. By some analysis Trump shouldn't be president if science were making itself clear.

   This difference between the ideal and the real often has some debaters citing the No-True-Scotsman "fallacy." Here is a description of that terminology.

• No True Scotsman
  

   Notice the problem there is the same as here. There is an "ideal" version of a "Scotsman" and a "real" one that doesn't measure up.  "Ideal" science might be godlike, but in reality scientists are human beings.

   As with many so called "fallacies" they are only a failure of logic in some circumstances and not in others. Sometimes it makes sense to appeal to popular opinion (ad populum), sometimes it makes sense to notice a speaker's flaws (ad hominem) and so on. Providing no other evidence is ignored and no faults misapplied both those can be fine. Here too it can make sense to discuss what science "ought" to be in some ideal should that topic arise. It could be a horrendous failure of logic though to assume that ideal presents itself in reality.

   It might be fair then to say that the defenders of science often commit the No-True-Scotsman "fallacy" as if "scientists" have no faith.

• A quick note on formal debate
  

   Discussing any selection of people ideal or real is fine, especially if it is a real group you have studied and wish to show your findings. Formal debate avoids the confusion by setting the definition of a "Scotsman" to whatever fits the immediate needs of the meeting, and is decided by mutual arbitrary agreement before the presentation of any arguments about those particular Scotsman. It is not illogical to define your terms. It is illogical not to. Of course you need to secure cooperation of the opposing team to discuss your particular definition.