The data being bad doesn't make conclusions drawn from it more trustworthy. Shouldn't it?
-
@topspin said in Tales from Coronavee-rooss Italy, mamma mia!:
@boomzilla read the part in the details section. The data being bad doesn't make conclusions drawn from it more trustworthy.
This is a problem for a lot of people.
So, how can an OMG devise a way such that
- making the data worse will
- make conclusions more trustworthy?
-
Doesn't
Eulervon Neumann whitening already do this?
-
I don’t understand. (
)
-
@topspin said in The data being bad doesn't make conclusions drawn from it more trustworthy. Shouldn't it?:
I don’t understand. (
)I don't understand why people are so mad that I even possibly misattributed the whitening algorithm. Was it Bresenham? I wanted to say Bresenham...
The goal of the thread is to break GIGO, or at least give it a sound thrashing.
-
@Gribnit said in The data being bad doesn't make conclusions drawn from it more trustworthy. Shouldn't it?:
The goal of the thread is to break GIGO, or at least give it a sound thrashing.
I'm not convinced you'll succeed, but there's plenty of G in here.
-
@Gribnit said in The data being bad doesn't make conclusions drawn from it more trustworthy. Shouldn't it?:
Doesn't Euler whitening already do this?
Euler? I hardly know 'er.
-
@Gribnit said in The data being bad doesn't make conclusions drawn from it more trustworthy. Shouldn't it?:
So, how can an OMG devise a way such that
- making the data worse will
- make conclusions more trustworthy?
Babbage had things to say about that.
-
@dkf said in The data being bad doesn't make conclusions drawn from it more trustworthy. Shouldn't it?:
@Gribnit said in The data being bad doesn't make conclusions drawn from it more trustworthy. Shouldn't it?:
So, how can an OMG devise a way such that
- making the data worse will
- make conclusions more trustworthy?
Babbage had things to say about that.
Indeed. And were he here he would have to stand informed. Consider, however, a biased but decorrelated bitstream being used to approximate
.Consider also
print 'No'.
-
@Gribnit is that an implementation of
betteridge.py?
-
@Gribnit said in The data being bad doesn't make conclusions drawn from it more trustworthy. Shouldn't it?:
devise a way such that
making the data worse will
make conclusions more trustworthy?Step 1: Discard data
Step 2: Print "This data was garbage".The worse the data, the more accurate the result. In fact, the opposite holds true too. Feed it good data, and the conclusion itself is very bad.
-
@cvi said in The data being bad doesn't make conclusions drawn from it more trustworthy. Shouldn't it?:
@Gribnit said in The data being bad doesn't make conclusions drawn from it more trustworthy. Shouldn't it?:
devise a way such that
making the data worse will
make conclusions more trustworthy?Step 1: Discard data
Step 2: Print "This data was garbage".The worse the data, the more accurate the result. In fact, the opposite holds true too. Feed it good data, and the conclusion itself is very bad.
<so>This should clearly be the accepted solution.</so>
-
@LaoC said in The data being bad doesn't make conclusions drawn from it more trustworthy. Shouldn't it?:
@cvi said in The data being bad doesn't make conclusions drawn from it more trustworthy. Shouldn't it?:
@Gribnit said in The data being bad doesn't make conclusions drawn from it more trustworthy. Shouldn't it?:
devise a way such that
making the data worse will
make conclusions more trustworthy?Step 1: Discard data
Step 2: Print "This data was garbage".The worse the data, the more accurate the result. In fact, the opposite holds true too. Feed it good data, and the conclusion itself is very bad.
<so>This should clearly be the accepted solution.</so>
Well, let's see what the data says about that...
Oh. Okay then.
