# What's the difference between stochastic and random?

What's the difference between stochastic and random?

• There is none. 
– Did
Commented Feb 28, 2012 at 9:10
• I don't like the term "random" because its vague and people misconstrue it as "evenly distributed", but I know of no technical difference. Commented Feb 28, 2012 at 9:13
• I agree with @AlexBecker. I would only add that random has many connotations (like entropy), not at all equivalent, and is a more generic term usable outside mathematics. Stochastic means nondeterministic or unpredictable. Random generally means unrecognizable, not adhering to a pattern. A random variable is also called a stochastic variable. Do random numbers exist? We speak of pseudorandom numbers. Commented Feb 28, 2012 at 9:57
• ... but I think that there is no a crucial difference in the meaning, only the difference in terminology used by different groups of scientists. I can say that also in Russia the equivalent of 'random' is used in old-style literature mostly, and 'stochastic' - in the modern one.
– SBF
Commented Feb 28, 2012 at 11:39
• @bgins Any example of occurrences of "stochastic variable", WP excepted?
– Did
Commented Aug 29, 2012 at 15:28

A variable is random. A process is stochastic. Apart from this difference, the two words are synonyms.

• As if random was a subproduct of stochastic? Commented Feb 28, 2012 at 16:58
• The term "stochastic variable" does occur sometimes. But more often "random" is used. Commented Feb 28, 2012 at 19:45
• I have seen the terminology "random process" used.
– user81375
Commented Jan 29, 2018 at 15:03
• I truly love the simple and intuitive way to explain the difference. Thank you. Commented Oct 10, 2020 at 21:08
• The current PyMC3 documentation uses the terms stochastic random variable and deterministic random variable to distinguish between those that have and those that don't have a parent in the dependency graph of random variables in a Bayesian model. Commented Apr 11, 2022 at 7:06

There is an anecdote about the notion of stochastic processes. They say that when Khinchin wrote his seminal paper "Correlation theory for stationary stochastic processes", this did not go well with Soviet authorities. The reason is that the notion of random process used by Khinchin contradicted dialectical materialism. In diamat, all processes in nature are characterized by deterministic development, transformation etc, so the phrase "random process" itself sounded paradoxically. Therefore, Khinchin had to change the name. After some search, he came up with the term stochastic, from στοχαστικὴ τέχνη, the Greek title of Ars conjectandi. Being popularized later by Feller and Doob, this became a standard notion in English and German literature.

Funny enough, in Russian literature the term "stochastic processes" did not live for long. The 1956 Russian translation of Doob's monograph by this name was already entitled Вероятностные процессы (probabilistic processes), and now the standard name is случайный процесс (random process).

• Very interesting! Is there a reference for the story? (it's OK if it's in Russian)
– Leo
Commented Jan 6, 2017 at 13:47
• @Leo, unfortunately, I can't recall the origin of the story. For a long time I had been thinking that it is from Shiryaev's book, Probability, but at the moment of writing this post I wasn't able to find it there, so didn't give a source. Commented Jan 6, 2017 at 14:56
• @Leo, by the way, there's a similar story (but with happy ending) about the notion of independence, which contradicted diamat much more seriously. This one is somewhere in Suhov, Kelbert Probability and Statistics by Example. Commented Jan 6, 2017 at 15:15
• So stochastic was a politically correct replacement for 'random'? Fascinating. Commented Nov 21, 2022 at 11:46

Neither word by itself has a commonly accepted formal definition in mathematics, so one cannot really ask about "the difference" between them.

They are used in phrases such as "random variable," "random walk," "stochastic process," "stochastically complete," etc, which have accepted definitions of their own. In all cases both words tend to refer to an element of chance or unpredictability. But they are generally not interchangeable; if you talk about a "stochastic walk" people will be confused.

Random process and stochastic process are completely interchangeable (at least in many books on the subject). Although once upon a time "stochastic" (process) generally meant things that are randomly changing over time (and not space). See relevant citations:

https://en.wikipedia.org/wiki/Stochastic_process#Terminology

In English the word "stochastic" is technical and most English speakers wouldn't know it, whereas, from my experience, many German speakers are more familiar with the word "Stochastik", which they use in school when studying probability.

The word "stochastic" ultimately comes from Greek, but it first gained its current sense, meaning "random", in German starting in 1917, when Sergei Bortkiewicz used it. Bortkiewicz had drawn inspiration from the book on probability by Jakob (or Jacques) Bernoulli, Ars Conjectandi. In the book, published 1713, Bernoulli used the phrase "Ars Conjectandi sive Stochastice", meaning the art of conjecturing. After being used in German, the word "stochastic" was later adopted into English by Joseph Doob in the 1930s, who cited a paper on stochastic procsses written in German by Aleksandr Khinchin.

https://en.wikipedia.org/wiki/Stochastic_process#Etymology

The use of the term "random process" pre-dates that of "stochastic process" by four or so decades.

Although in English the word "random" does come from French, I strongly doubt it ever meant random in French. In fact, it originally was a noun in English meaning something like "great speed". It's related to the French word "randonée" (meaning hike or trek), which is still used today. To describe a random variable, French uses the word "aléatoire", stemming from the Latin word for dice (which features in a famous quote "Alea iacta est." by Julius Caesar). The English equivalent "aleatory" is not commonly used (at least in my random circles).

• while the word "randonnée" was derived from the now disused word "randon", the meaning of "randon" was "impetuous, haphazard". (A "randonnée" is a disorganized hike, you could almost say: a "random walk") Quote of the day: "L'hiver survint avec grande furie/ Monceaux de neige et grands randons de pluie" (Jean de la Fontaine). Commented Nov 24, 2022 at 4:34

Stochastic comes from Ancient Greek whereas random is an old French word. (fun fact: random has totally disappeared in modern French and was replaced by aléatoire which comes from... Latin)

Otherwise there is no difference between them in the realm of Probability Theory.

The term stochastic in Hydrology science refers to a process which periodically and apparently-independently happens but a kind of dependency exists. For example, if the flow of a river in last (say) 2 weeks has been low, it will probably be low in the next weeks too. So, the flow of a river is not a complete random variable but stochastic.

• Likewise, I've noticed that network theory tends to refer to traffic as being stochastic. Such traffic, from the point of view of a router, would be considered random, but of course each packet was deterministically produced. Commented Sep 4, 2014 at 18:15

In Chinese literature, there is no difference between those two terms at all.

Both of the "stochastic" or "random" are 随机 in Chinese.

Thus, I would argue that the use of "stochastic" and "random" does not differ in mathematics, but only in language conventions.

The terms "stochastic variable" and "random variable" both occur in the literature and are synonymous. The latter is seen more often. Similarly "stochastic process" and "random process", but the former is seen more often.

Some mathematicians seem to use "random" when they mean uniformly distributed, but probabilists and statisticians don't. I suspect those who do that haven't thought about it much.

• Any example of occurrences of "stochastic variable", WP excepted?
– Did
Commented Aug 29, 2012 at 15:28
• @did : I don't have any at hand, but I've seen it in print. Commented Aug 29, 2012 at 16:28
• @did : google.com/… Commented Aug 29, 2012 at 16:30
• scholar.google.com/… Commented Aug 29, 2012 at 16:31
• Thanks for the links. After skimming very partly through them, what strikes me is that their majority is related to applications of mathematics (electrical engineering, management sciences, econometrics, physics, artificial intelligence, automatics, water resources research, others) rather than to mathematics and/or probability theory per se. My guess is that the frequency of "stochastic variable" would vanish, or nearly so, if the corpus was restricted to these fields.
– Did
Commented Aug 29, 2012 at 17:18

A random process is unpredictable such as the movement of the tip of a feather In wind. If we assume that the movement of a roller coaster is deterministic, then a stochastic process would be the movement of the tip of a feather attached to a moving roller coaster. That is to say, stochastic processes have components that are both deterministic AND random; e.g. Martingales.

So. There are Random ODEs and Random PDEs that are not synonymous to SODEs and SPDEs -- although wide classes of RODEs can be converted to SODEs (usually replacing random normal forcings for Ornstein-Uhlenbeck dynamics) through the Imkelller-Schmallflusss correspondence.

I don't know that much about the theory because my work is in simulating/numerically solving these. But they're a natural formulation for systems such as earthquakes, tumor growth, etc.

A nice wide-ranging textbook aimed at scientists (spends a third of the book introducing rigorous probability before entering the subject) is

NECKEL, Tobias; RUPP, Florian. Random Differential Equations in Scientific Computing. Walter de Gruyter, 2013.

A textbook of numerics, a little narrower in scope is

Han, Xiaoying, and Peter E. Kloeden. "Random Ordinary Differential Equations and Their Numerical Solution." (2017).

Kloeden is well known for his textbook on numerical SDEs, of course. If you plug these on Google Scholar you can browse more recent papers that cite them for hours.

EDIT: I found this slideset from a talk by Neckel [PDF link]. It defines RODEs and explains the intuition for the Imkeller-Schmallfluss correspondence.

An extreme example of a stochastic process is a deterministic signal X(t)=f(t). In this case, E{X(t)}=E{f(t)}=f(t), R(t1,t2)=E{f(t1)f(t2)}=f(t1)f(t2)

AS another example Using the Poisson points t_i, we form a process x(t) such that x(t)=1 if the number of points in the interval (0, t) is even, and x(t)=-1 if this number is odd. E{x(t)} = exp{-2landa.t}, R(t1,t2) = exp{-2landa.|t1-t2|}

This process is called semi-random telegraph signal because its value x(0)=1 at t=0 is not random. To remove this certainty, we form the product y(t)=ax(t) where a is a random variable taking the values +1 and -1 with equal probability and is independent of x(t). The process y(t) so formed is called random telegraph signal. Since E(a)=0 and E(a^2)=1, the mean of y(t) equals E{a}E{x(t)}=0 and its auto-correlation is given by E{y(t1)y(t2)} = E{a^2}E{X(t1)X(t2)} = exp{-2landa.|t1-t2|}.

Reference: Papoulis Edition 4.

• I need a reference regarding telegraph processes. Could I ask you to look into this question? Commented Jul 31, 2022 at 20:27
• @samwolfe In addition to 1- Papoulis Edition 4, there are some other books that will be mentioned in the following, but I am not sure whether they can be helpful or not. : 2-"Probability and Random Processes, Second Edition With Applications to Signal Processing and Communications by Scott Miller, Donald Childers", 3-"Probability Theory and Stochastic Processes by Pierre Brémaud", 4-"Probability, Statistics, and Random Processes for Engineers by Henry Stark, John W. Woods". Commented Aug 31, 2022 at 8:21

From a remote sensing point of view, we usually refer to a bounded but unpredictable process as stochastic. If the process were unbounded and unpredictable I would tend to use random, but this case doesn't occur very much in my world! :)

I believe there is a difference between random and stochastic. Random has no preciptating or a priori cause i.e acausal. Random action stands alone - not within any system. Stochastic is random, but within a probablistic system. In other words an act of God is random, but a hurricane hitting the east coast of the US is stochastic event. Any individual hurricane may be random but it also exists as a mathmatical probability within a system of many hurricanes that hit or do not hit the east coast every year. The later is therefore stochastic. A coin flip has an interesting difference than a hurricane since each individual flip of a coin is already stochastically limited to the 50% statistical probability of two possible results.

I will quote from Robert Gray Gallager's MIT OpenCourseWare notes for "Discrete Stochastic Processes" (1):

"Stochastic and random are synonyms, but random has become more popular for rv’s (random variable) and stochastic for stochastic processes. The reason for the author’s choice is that the common-sense intuition associated with randomness appears more important than mathematical precision in reasoning about rv’s, whereas for stochastic processes, common-sense intuition causes confusion much more frequently than with rv’s. The less familiar word stochastic warns the reader to be more careful." (Chapter 1: Introduction and review of probability, page 15, fn 15).

I would make a distinction: for example in a queueing system the arrival times (or interval times) might be modelled by a Poisson process which would be time independent and would not be bound my initial conditions. This would be an example of a random process which outputs random variables. Service time in the queue would be dependent on the previous state(s) of the system and possibly initial conditions. This would be an example of a stochastic process which also outputs random variables. I am not sure that the term ‘stochastic variable’ has any real meaning except possibly to indicate how the variable was produced.

• As this is a four year old Question with an Accepted Answer (and others), Readers would benefit from your supporting any new material in your Answer with some references to the literature, etc. Commented Jul 23, 2016 at 1:31

There is absolutely a difference between stochastic process and randomness. For example, if I take one step then let's suppose my friend takes two steps. Now my friend's steps are not random, those are dependent on my steps. That means my friend's steps has a process which is the number of steps I take. But it is random because my friend doesn't know how many steps I will take. So the steps I take is a random walk.

• // , Would you please make this example more clear, perhaps by listing its assumptions first, e.g. "Assuming I take a random walk, and let us assume I have a friend, who..." Commented Jan 29, 2017 at 2:19