Rugxdoma:

...: the first occurrence of probable in the Tekstaro is in 1978 (John Wells) ...

I think this late date reflects the fact that the word "probability" became widely used in some languages under the influence of the "New Maths" school curriculum, which was developed in the United States, when challenced by the Soviet Union's space programme. This "Sputnic shock" maths, composed mainly of set theory, probability theory and vector calculation, was shortly afterwards introduced in many countries.
Sweden as well as Finland adopted it for a period of time, but never changeed the traditional name of "probability".
If you translate "probable" into Swedish, it will be "sannolik". This word is composed of two parts: "sann", which means thrue, and likhet, which means "alike" or "similar", thus exactly as "versxajna" is. It is used in all registers, always. So I do not see that there should be any difference in Esperanto either between the two words.

I second your opinion. I suppose that Zamenhof took "versxajna" from the German "wahrscheinlich", which (like "sannolik" ) means wahr-schein-lich = ver-sxajn-a...
Also in German, "probability calculus" is "Wahrscheinlichkeitsrechnung".

If you translate "probable" into Swedish, it will be "sannolik". This word is composed of two parts: "sann", which means thrue, and lik, which means "alike" or "similar", thus exactly as "versxajna" is. It is used in all registers, always. So I do not see that there should be any difference in Esperanto either between the two words.

With respect Rugxdoma (and Nornen), I don't think the argument that a samforma vorto in some natural language or languages has a certain range of applications means that the Esperanto word's semantic range should be the same.

It is fairly commonplace that the cognate Esperanto word has a more restricted range than in the source language or languages.

You might think the distinction I see between probable and verŝajne is not worth making. That's a point of view. But as a general principle it is better that Esperanto be able to make distinctions than be cluttered with exact synonyms.

It is the same with french:
vraisemblable = vrai + semblable
vera + simila / sxajna

What about "malversxajna" ? (in french, invraisemblable, whose meaning is close to "nekredebla")
In tatoeba, I found this:

http://tatoeba.org/fre/sentences/show/540628

Lia sukceso estas malverŝajna.

What is the meaning of the above sentence?
(a) He certainly will not succeed.
(b) He succeeded, but everybody thought he would not.

sudanglo:You might think the distinction I see between probable and verŝajne is not worth making. That's a point of view. But as a general principle it is better that Esperanto be able to make distinctions than be cluttered with exact synonyms.

I cannot see the point in starting seeing differences where there are no differences. Either will these subtleties not be understood or you have to convince all others to see what you see. Cluttered has the languages been already when one unnecessary word was imported. Claiming that the two are different only protracts the process by which we get rid of one of them.