# It's not “probability” but…? [closed]

I have values between 0 and 1, like [0.9, 0.8,...], indicating that a value closer to 1 is more probable than one close to 0, but without all the values adding up to 1. I guess I can't call it probability, but what else can I call those values?

-

## closed as not a real question by MετάEd, tchrist, Daniel, Mitch, JLGOct 12 '12 at 3:48

It's difficult to tell what is being asked here. This question is ambiguous, vague, incomplete, overly broad, or rhetorical and cannot be reasonably answered in its current form. For help clarifying this question so that it can be reopened, visit the help center.If this question can be reworded to fit the rules in the help center, please edit the question.

I think you need to supply more context. What are these numbers measuring, exactly? – user16269 Sep 8 '12 at 17:40
Why aren't these "probabilities"? That sounds like the right word from what you've written here. – SevenSidedDie Sep 8 '12 at 17:49
@SevenSidedDie, the values of discrete probabilities that apply to a situation total to 1; likewise for the integral of a probability distribution. In the case at hand, one might refer to relative probabilities, or could make up terms like probability index or likelihood index. – jwpat7 Sep 8 '12 at 18:34
@jwpat7 These could easily be called probabilities, depending on context. "Probability of A happening: 0.9; Probability of B happening: 0.8". It depends on whether these are independent or exclusive possible things, which is the context that we're lacking. SF's "bias" is a good general-purpose word that smooths over the differences between possible contexts that I can imagine. – SevenSidedDie Sep 8 '12 at 18:42
@SevenSidedDie, I agree context is lacking and of course that non-independent probabilities can total more than 1 (also, by “discrete probabilities that apply to a situation” I meant the independent probabilities that describe all the independent events possible) but disagree that bias is a suitable word. In statistics, bias has nothing to do with relative probabilities and is a wrong or misleading word in that context. – jwpat7 Sep 8 '12 at 19:01

You mean, an event with "0.9" doesn't actually have a 9 in 10 chance for success, but it's more probable than the one with "0.8"?

I guess you could call it "Bias". This is what's used to indicate tendency towards given option/outcome/point while not stating the value precisely, so the event with bias 0.9 is more biased towards success than the one with bias of 0.8.

You may also talk about unnormalized probability, where the values do add up to a certain constant but not to 1 and proportionally correspond to relative probabilities - you'd need to normalize them by dividing each by their sum. But if no such easy transformation is possible, "bias" is the safe choice.

-

I'm risking retribution by not answering your actual question, but if SF in his answer has deduced what you're actually getting at, I suggest that you shy away from using a scale from 0 to 1 to show 'pseudo-probabilities' - it's just going to produce confusion by adapting an agreed convention in a non-standard way. Probabilities involve ratio data not merely ordinal data. ( http://www.usablestats.com/lessons/noir )

Why not use instead a Likert-like scale ( http://en.wikipedia.org/wiki/Likert_scale ):

A = certain to happen B = very probably going to happen ...

Z = certain not to happen (use as many graduations as you think sensible).

This will almost certainly be subjective (ie not totally accurate) but is only a reformulation of your own idea, without the confusion of terminology. Yes, someone will ask what a 'C' classification say actually means, but that's fair enough and much fairer than leaving some people believing that a pseudo-probability of 0.8 means a probability of 0.8 :-/

-