For Chopin, I vote for the Paderewski Edition. Paderewski worked closely with the scholars L. Bronarski and J. Turczynski. It is well researched and the end-of-volume editorial commentaries are highly informative in comparing the autograph and original editions in France, England and Germany.
So, for example, on your question as to why the B octave in the LH in Alfred versus the G octave, the Paderewski editorial commentary tells you that the B is found in the autographed manuscript preserved at the Frederic Chopin Society in Warsaw, as well as in the first German edition (Breitkopf & Hartel) which was based on that autographed manuscript. The implication is that the G originated in the original French edition (M. Schlesinger in Paris). Both notations were subsequently picked up by still other publishers over time. Thus, there has developed two different performance practices on the B versus the G.
You'll find these note discrepancies very frequently in Chopin. A familiar example is the choice of and E flat or E in the Prelude, Op. 28, No. 20 in Cm, top of the last RH chord at the end of the third measure. Half of pianists play it one way, and the other half the other way. In recitals knowing listeners will have their ears cocked to hear which which note the pianist will play. So before questioning a "wrong note" in Chopin, you really have to first ask about the edition (as you have done here), as the performer might well have played exactly what is printed in that edition.