Guessing is the act of drawing a swift conclusion, called a guess, from data directly at hand, which is then held as probable or tentative, while the person making the guess (the guesser) admittedly lacks material for a greater degree of certainty.

A guess is an unstable answer, as it is "always putative, fallible, open to further revision and interpretation, and validated against the horizon of possible meanings by showing that one interpretation is more probable than another in light of what we already know". In many of its uses, "the meaning of guessing is assumed as implicitly understood", and the term is therefore often used without being meticulously defined.

Guessing may combine elements of deduction, induction, abduction, and the purely random selection of one choice from a set of given options. Guessing may also involve the intuition of the guesser, who may have a "gut feeling" about which answer is correct without necessarily being able to articulate a reason for having this feeling.

Philosopher Mark Tschaepe, who has written extensively on the scientific and epistemological role of guessing, has noted that there are often-overlooked "gradations" of guessing — that is, different kinds of guesses susceptible to different levels of confidence. Tschaepe defines guessing as "an initial, deliberate originary activity of imaginatively creating, selecting, or dismissing potential solutions to problems or answers to questions as a volitional response to those problems or questions when insufficient information is available to make merely a deduction and/or induction to the solution or answer". He objects to definitions that describe guessing as either forming a "random or insufficiently formed opinion", which Tschaepe deems too ambiguous to be helpful, or "to instantaneously happen upon an opinion without reasoning". Tschaepe notes that in the latter case, the guess might appear to occur without reasoning, when in fact a reasoning process may be occurring so quickly in the mind of the guesser that it does not register as a process. This reflects the observation made centuries before by Gottfried Wilhelm Leibniz, that "when I turn one way rather than another, it is often because of a series of tiny impressions of which I am not aware". Tschaepe quotes the description given by William Whewell, who says that this process "goes on so rapidly that we cannot trace it in its successive steps".

A guess that "is merely a hunch or is groundless... is arbitrary and of little consequence epistemologically". A guess made with no factual basis for its correctness may be called a wild guess. Jonathan Baron has said that "[t]he value of a wild guess is l/N + l/N - l/N = l/N", meaning that taking a true wild guess is no different from choosing an answer at random. Philosopher David Stove described this process as follows:

A paradigm case of guessing is, when captains toss a coin to start a cricket match, and one of them 'calls', say "heads". This cannot be a case of knowledge, scientific knowledge or any other, if it is a case of guessing. If the captain knows that the coin will fall heads, it is just logically impossible for him also to guess that it will. More than that, however: guessing, at least in such a paradigm case, does not even belong on what may be called the epistemic scale. That is, if the captain, when he calls "heads", is guessing, he is not, in virtue of that, believing, or inclining to think, or conjecturing, or anything of that sort, that the coin will fall heads. And in fact, of course, he normally is not doing any of these things when he guesses. He just calls. And this is guessing, whatever else is.

In such an instance, there not only is no reason for favoring "heads" or "tails", but everyone knows this to be the case. Tschaepe also addresses the guess made in a coin flip, contending that it merely represents an extremely limited case of guessing a random number. Tschaepe examines such guesses at greater length with the instance of guessing a number between 1 and 100, for which Tschaepe notes that the guesser "has to look for clues that are specific to what or whom is ordering them to guess, as well as possible past scenarios that involved guessing numbers", and once these are exhausted, "there comes a point very early in the process wherein no other clue to an answer exists". As an exemplary case of guessing that involves progressively more information from which to make a further guess, Tschaepe notes the game of Twenty Questions, which he describes as "similar to guessing a number that the other person is thinking, but unlike guessing a number as a singular action... allows for combining abductive reasoning with deductive and inductive reasoning".

An apparently unreasoned guess that turns out to be correct may be called a happy guess, or a lucky guess, and it has been argued that "a 'lucky guess' is a paradigm case of a belief that does not count as knowledge". Jane Austen, in Emma, has the titular character respond to a character calling a match that she made a "lucky guess" by saying that "a lucky guess is never merely luck. There is always some talent in it". As Tschaepe notes, William Whewell stated that certain scientific discoveries "are not improperly described as happy Guesses; and that Guesses, in these as in other instances, imply various suppositions made, of which some one turns out to be the right one".