3 ms·

Extraordinary claims require extraordinary proof. When it's mostly likely experimental error, why do you think anyone should "take it for an anomaly"?

golergka 10y ago
Extraordinary claims require extraordinary proof to take them as truth. Less extraordinary proof can still let them be a "may be". That's all they're trying to be for now.
- greglindahl 10y ago
  I was responding to someone who said "it's wrong of mainstream science immediately reject Cold Fusion/LENR or even Em Drive research". In response to you, "unlikely" is not the same as "may be".
  - snowwrestler 10y ago
    It not good science to reject a hypothesis or anomalous result before the data makes clear what is going on. Imagine if the guys at Bell Labs had just said "eh, who knows, nevermind" when they couldn't figure out the hiss in the horn antenna.
    - greglindahl 10y ago
      There was no theory saying that the horn antenna should have no hiss. It was a classic example of being the first measurement of a thing. It happens all the time in astronomy. Since this measurement people have been much more open to whatever the data says when they observe in a new frequency band. If anything, the surprise in this particular case was mainly because the people doing the experiment weren't astronomers, and had never considered what they might see if they looked up. Excellent discovery, and easy and inexpensive to replicate.
      - snowwrestler 10y ago
        > There was no theory saying that the horn antenna should have no hiss. Then why did they go to such extraordinary lengths to find the source of it? Why didn't they just shrug and work around it? Keep in mind they initially had no knowledge of the theoretical prediction of CMB, which was fairly recent and controversial. They just had a weird observation and busted their butts trying to figure out what it was. > It was a classic example of being the first measurement of a thing. Yes, now we know that. At the time it was not clear that it was a phenomenon worth measuring at all. I mean, they thought it was pigeons. It looked like a dumb error at first. > Excellent discovery, and easy and inexpensive to replicate. The "emdrive" is also easy and cheap to replicate. I'm not saying that the emdrive works, I'm saying that experimental results matter. They should not be dismissed just becaus me they seem to violate theory. Violating theory is why people experiment in the first place.
        greglindahl 10y ago
        Well, they were mostly weirded out that it was the same hiss no matter where in the sky they pointed the horn, and indeed, that's pretty weird. It was non-obvious, and now we know why. For the emdrive, the effect is so small that it's suggested that the next step is to launch a powerful satellite costing $100mm or more. Uh, really? From your comment, you seem to be tangled up in results being "dismissed". That's not what's going on, when people make comments about extraordinary claims and extraordinary results. A tiny, hard to measure result is not an OH MY GOD THE THEORY IS WRONG result. It is likely an experimental error.
        StanislavPetrov 10y ago
        Its ironic that skeptics of the EM drive not only refuse to consider the actual, empirical evidence, but also oppose testing it in the only way that is beyond dispute.
        beevai142 10y ago
        It has already been independently tested (TU Dresden, NWPU). They saw only noise.
  - AnimalMuppet 10y ago
    "May be" covers everything from "unlikely, but not quite impossible" to "almost but not quite certainly true". A couple of times in this discussion, you've argued about the definitions of words, and your definitions have been... at least different from the way the words are normally used. It's kind of distracting from the actual discussion, without adding much.
    - greglindahl 10y ago
      Well, sorry, I'm trained in physics. It's the way we use those words.
      - AnimalMuppet 10y ago
        Hmm. I'm trained in physics, too, to some degree, but I stopped at a bachelor's degree. What, in your usage, is the difference between "unlikely" and "may be"? (By the way, I'm not arguing against "unlikely" wrt any of these things...)
        greglindahl 10y ago
        "may be" doesn't give you an idea if it's 50% or 1% or 0.0001% likely. "Unlikely" does give you an idea: the probability is small. When you're talking about an extraordinary claim, it's a good idea to use "unlikely" and not "may be". The moon "may be" has deposits of green cheese, we just haven't found them yet. It is unlikely that the moon has deposits of green cheese. It is likely that the EM Drive thrust measurement is measurement error.
    - dekhn 10y ago
      Science writing, especially the hard core sciences, has a collection of terms that are used among practitioners and have very specific meanings that are well-understood. These usages convey fairly subtles signals about one's understanding of, and confidence in, a physical phenomenon. This is typically only learned by people who write papers and have advisors who can guide them. A really nice example of this which is also immediately readable is Watson and Crick's 1953 DNA structure paper. http://www.nature.com/nature/dna50/watsoncrick.pdf http://www.nature.com/nature/dna50/watsoncrick.pdf It's a classic understated style, with a small amount of boastful brilliance. The passive tone is a bit dated, but you can clearly see little was known about how DNA worked, and they were very careful not to make any strong claims without qualifiers.
- dekhn 10y ago
  Actually extraordinary claims require no more proof than any other. Otherwise, you're prone to accepting ordinary claims with less evidence.
  - CoryG89 10y ago
    You don't need more proof, just extraordinary proof. Totally different than regular proof. :P
  - tbrownaw 10y ago
    Well, yes. That's how priors are supposed to work. Something that's 1% likely and something that's 75% likely require different amounts of evidence to get to 95%.
onion2k 10y ago
You don't often go from "nothing" to "proof it works" in a single iteration of a new technology.
- greglindahl 10y ago
  Indeed, if you look at the history of science, even modern examples such as the discovery of H. Pylori, it's a messy process. However, don't be confused by survivor bias: almost all extraordinary results turn out to be experimental error.
  - snowwrestler 10y ago
    Fine, but let's also recognize that this is not yet one of them. It might be an error, in fact it almost certainly will be. But it is not yet certainly an error. Stand back: science still in progress!
- Retric 10y ago
  But you should not be trying to prove it works. You should be trying to disprove it, until no other reasonable explanation is left. The most obvious example is to use normal water instead of heavy water and see if you get similar results.
Everhusk 10y ago
Agreed. However, rather than ignoring the experiment, I think it would be better to encourage scientists to understand why. “Whatever Nature has in store for mankind, unpleasant as it may be, men must accept, for ignorance is never better than knowledge.” ― Enrico Fermi
- greglindahl 10y ago
  I didn't suggest ignoring the experiment. I'm just setting a baseline for how important this result is.
  - AnimalMuppet 10y ago
    Well, if it's experimental error, then we learn a bit more about how to properly and carefully conduct experiments. If it's not experimental error, though... Unfortunately, you have to look into it carefully to find out whether it's experimental error or something more interesting. And, to return to Everhusk's first post: Experimental error can result in an "anomaly". The experiment is not returning the results we expect. That's an anomaly. An anomaly doesn't have to be new physics.
jcoffland 10y ago
But you cannot get to extraordinary proof of you immediately reject the idea.
- greglindahl 10y ago
  Um, I didn't immediately reject the idea.
  - jcoffland 10y ago
    https://en.wikipedia.org/wiki/You#Informal_plural_forms https://en.wikipedia.org/wiki/You#Informal_plural_forms

golergka 10y ago

Extraordinary claims require extraordinary proof to take them as truth. Less extraordinary proof can still let them be a "may be". That's all they're trying to be for now.

greglindahl 10y ago
I was responding to someone who said "it's wrong of mainstream science immediately reject Cold Fusion/LENR or even Em Drive research". In response to you, "unlikely" is not the same as "may be".
- snowwrestler 10y ago
  It not good science to reject a hypothesis or anomalous result before the data makes clear what is going on. Imagine if the guys at Bell Labs had just said "eh, who knows, nevermind" when they couldn't figure out the hiss in the horn antenna.
  - greglindahl 10y ago
    There was no theory saying that the horn antenna should have no hiss. It was a classic example of being the first measurement of a thing. It happens all the time in astronomy. Since this measurement people have been much more open to whatever the data says when they observe in a new frequency band. If anything, the surprise in this particular case was mainly because the people doing the experiment weren't astronomers, and had never considered what they might see if they looked up. Excellent discovery, and easy and inexpensive to replicate.
    - snowwrestler 10y ago
      > There was no theory saying that the horn antenna should have no hiss. Then why did they go to such extraordinary lengths to find the source of it? Why didn't they just shrug and work around it? Keep in mind they initially had no knowledge of the theoretical prediction of CMB, which was fairly recent and controversial. They just had a weird observation and busted their butts trying to figure out what it was. > It was a classic example of being the first measurement of a thing. Yes, now we know that. At the time it was not clear that it was a phenomenon worth measuring at all. I mean, they thought it was pigeons. It looked like a dumb error at first. > Excellent discovery, and easy and inexpensive to replicate. The "emdrive" is also easy and cheap to replicate. I'm not saying that the emdrive works, I'm saying that experimental results matter. They should not be dismissed just becaus me they seem to violate theory. Violating theory is why people experiment in the first place.
      - greglindahl 10y ago
        Well, they were mostly weirded out that it was the same hiss no matter where in the sky they pointed the horn, and indeed, that's pretty weird. It was non-obvious, and now we know why. For the emdrive, the effect is so small that it's suggested that the next step is to launch a powerful satellite costing $100mm or more. Uh, really? From your comment, you seem to be tangled up in results being "dismissed". That's not what's going on, when people make comments about extraordinary claims and extraordinary results. A tiny, hard to measure result is not an OH MY GOD THE THEORY IS WRONG result. It is likely an experimental error.
        StanislavPetrov 10y ago
        Its ironic that skeptics of the EM drive not only refuse to consider the actual, empirical evidence, but also oppose testing it in the only way that is beyond dispute.
        beevai142 10y ago
        It has already been independently tested (TU Dresden, NWPU). They saw only noise.
- AnimalMuppet 10y ago
  "May be" covers everything from "unlikely, but not quite impossible" to "almost but not quite certainly true". A couple of times in this discussion, you've argued about the definitions of words, and your definitions have been... at least different from the way the words are normally used. It's kind of distracting from the actual discussion, without adding much.
  - greglindahl 10y ago
    Well, sorry, I'm trained in physics. It's the way we use those words.
    - AnimalMuppet 10y ago
      Hmm. I'm trained in physics, too, to some degree, but I stopped at a bachelor's degree. What, in your usage, is the difference between "unlikely" and "may be"? (By the way, I'm not arguing against "unlikely" wrt any of these things...)
      - greglindahl 10y ago
        "may be" doesn't give you an idea if it's 50% or 1% or 0.0001% likely. "Unlikely" does give you an idea: the probability is small. When you're talking about an extraordinary claim, it's a good idea to use "unlikely" and not "may be". The moon "may be" has deposits of green cheese, we just haven't found them yet. It is unlikely that the moon has deposits of green cheese. It is likely that the EM Drive thrust measurement is measurement error.
  - dekhn 10y ago
    Science writing, especially the hard core sciences, has a collection of terms that are used among practitioners and have very specific meanings that are well-understood. These usages convey fairly subtles signals about one's understanding of, and confidence in, a physical phenomenon. This is typically only learned by people who write papers and have advisors who can guide them. A really nice example of this which is also immediately readable is Watson and Crick's 1953 DNA structure paper. http://www.nature.com/nature/dna50/watsoncrick.pdf http://www.nature.com/nature/dna50/watsoncrick.pdf It's a classic understated style, with a small amount of boastful brilliance. The passive tone is a bit dated, but you can clearly see little was known about how DNA worked, and they were very careful not to make any strong claims without qualifiers.
dekhn 10y ago
Actually extraordinary claims require no more proof than any other. Otherwise, you're prone to accepting ordinary claims with less evidence.
- CoryG89 10y ago
  You don't need more proof, just extraordinary proof. Totally different than regular proof. :P
- tbrownaw 10y ago
  Well, yes. That's how priors are supposed to work. Something that's 1% likely and something that's 75% likely require different amounts of evidence to get to 95%.

onion2k 10y ago

You don't often go from "nothing" to "proof it works" in a single iteration of a new technology.

greglindahl 10y ago
Indeed, if you look at the history of science, even modern examples such as the discovery of H. Pylori, it's a messy process. However, don't be confused by survivor bias: almost all extraordinary results turn out to be experimental error.
- snowwrestler 10y ago
  Fine, but let's also recognize that this is not yet one of them. It might be an error, in fact it almost certainly will be. But it is not yet certainly an error. Stand back: science still in progress!
Retric 10y ago
But you should not be trying to prove it works. You should be trying to disprove it, until no other reasonable explanation is left. The most obvious example is to use normal water instead of heavy water and see if you get similar results.

Everhusk 10y ago

Agreed. However, rather than ignoring the experiment, I think it would be better to encourage scientists to understand why. “Whatever Nature has in store for mankind, unpleasant as it may be, men must accept, for ignorance is never better than knowledge.” ― Enrico Fermi

greglindahl 10y ago
I didn't suggest ignoring the experiment. I'm just setting a baseline for how important this result is.
- AnimalMuppet 10y ago
  Well, if it's experimental error, then we learn a bit more about how to properly and carefully conduct experiments. If it's not experimental error, though... Unfortunately, you have to look into it carefully to find out whether it's experimental error or something more interesting. And, to return to Everhusk's first post: Experimental error can result in an "anomaly". The experiment is not returning the results we expect. That's an anomaly. An anomaly doesn't have to be new physics.

jcoffland 10y ago

But you cannot get to extraordinary proof of you immediately reject the idea.

greglindahl 10y ago
Um, I didn't immediately reject the idea.
- jcoffland 10y ago
  https://en.wikipedia.org/wiki/You#Informal_plural_forms https://en.wikipedia.org/wiki/You#Informal_plural_forms