I envy your eloquence..

I concur: a fine introduction to the game narrative. The Russians used to do this sort of thing - probably still do - and there is a similar treatment in Bobby Fischer's "60 Memorable Games".

mmm the exchange variation.... it has a drawish reputation......

Very drawish, which might make it a good option against a stronger opponent. It won't give you many winning chances, but the 'threat' of taking out what Jo calls a 'High Draw', puts a strong Black player on his mettle. Even had White lost this game, it would have been very instructive in how to extract the most from slender advantages.

dxc4 is slightly unusual, aiding white's development.....

Trying to unbalance the game perhaps? Black might be 'giving odds' a little here in order to create dynamic imbalances that will give chances for Black to turn the tables. Such a policy can yield dividends, but does carry risks.

yeah, it does seem to expose White's king a little bit... for example, there's the open diagonal for Black's bishop now

Very weakening, and no doubt Black was happy to see this. Now we must see whether Black has the resources to exploit that weakening, or withal, whether White's piece activity will compensate. It is important to remamber, too, what might be available to the bishop once it reaches g6. There doesn't seem to be all that much on the central light squares...

but following the recommended move in the database....

Threatens to 'Noah's Ark' Black's bishop. The thing about pawn storms of the kind we are looking at is that they work much better with the centre closed up - or with the potential to become closed. That doesn't prevail here.

and following Boersma, Paulus A (2350) vs. Van der Mije, Alexandra (2300) 1979 in the database

we seem to be out of the databases now....

Rxe5 also avoids completely destroying White's pawn structure (isolated pawns on the e, c and a files? no thanks!)

The resulting position looks very promising for White: a central majority, greater space, a lead in development, and the bishop pair vs a knight pair on an open board...

hm... my first thought was to Qd3: it looks to me that you gain a tempo considering the development of the last rook. And I don't see counter move.

I like 19.Qd3 as well, but I can see here that Jo was combining attack with defence: holding g3, and doubling major pieces on the e-file. Good thinking. Unfortunately, Black can at once challenge White's possession of the file.

here is the tempo lost.

I have a feeling there ought to be a better spot for that bishop. Intriguing is 22.Bf4, with this alluring possibility in mind: 22...Qa5!? 23.Re1 Qxc3?? 24.Rxe8 Nxe8 25.Bxf7ch! Kxf7?? 26.Be2+ etc. The c-pawn thus insured against death suggests that this more advanced posting might have been preferred for the bishop..

I would have thought with pawn on both sides of the board, three pawn islands each, open board, the two bishops should be slightly better than the two knights....

yeah, if White can get to an endgame...

of course two knights are potentially dangerous, but it is easier to open diagonals than to walk the lands with horses...

At first sight the two bishops ought to confer upon White an edge in this late middle game/early endgame. But Black's knights are quite active and well placed to cooperate with each other. White's dark square bishop is the problem child - confined to a rather dreary spot at b2. He cries out for open spaces... I think the presence of the Queens have also to be borne in mind. Generally speaking, Q-and-N work together rather better than Q-and-B, which means that White has to be very circumspect about exchanging a minor piece. At the same time, Black is probably better off keeping the Queens on. The propects are very intriguing.

I would say to prevent Nxh3+. If not, the move here would have been g5, winning a piece due to the back rank mate threat. Even after h5, the threat is still alive, since g5, Nd5, Qe8+, Kh7, Bc2+, g6, and c5-d6 should win. BUT for the Qc3+ / Nxh3+ moves. Is there no way to make it work? I don't know, maybe Kh1, that delayes black checks just one tempo...

can't black play Nd7 and Nf8?

Yes, but... 25.g5 Nd7?* 26.Qe8+ Nf8 27.Ba3 Ng6 28.Bd6 Qd7 (28....Qb7 29.h4) 29.Qxd7 Nxd7 30.Ba4 wins the c-pawn. Of course 25...Nxh3+ really puts the kibosh on 25.g5?? Jo's choice is best. (Not, of course, the horrible 25.Qg3??)

I would have thought the bishop was slightly better. If anything surely it is the better to the two bishops, the Bc1 is slightly hindered by the c3 and d4 pawns.

Ah here you made a positional error. 27. Bxd5 is to greedy and the exchange was not really on your terms the knight could have proved very useful. so what would i have done 27. Qe8 only move for black Kh7 28. Kh1 now black can take with 28...hxg4 resulting in 29. Qe4+ Ng6 30. Qxg4 Or the better move 28...g6 resulting in 29. c4 Nc3 30. gxh5 Nxh5 ( diagram) both variation give a better positional play for white ( therefore better chances )...

I agree with Freddy (easy19) on this. The Q check on e8 offers some interesting possibilities. I was thinking more in terms of 27.Qe8ch Kh7 28.Qe4ch. If now, 28...Kg8, White can at once suggest a draw by 29.Qe8+, or try for more by 29.Bxf4! [a] 29...Nxf4 30.Qe8+ Kh7 31. Bxf7! wins. Black can wriggle, but not escape: 31...Nxh3+ 32.Kg2 Nf4+ 33.Kf3 hxg4+ 34.Kxg4. Better for Black is [b] 29...Qxf4 30.Bxd5 Qg3 with a likely draw. Instead of 28...Kg8, Black can try ...g6 or ...Ng6. White is better advised not to take 29.gxh4 in either case, but combine defence with attack by retiring the Q to f3.

The game continuation is very like one of the lines I suggested in my previous comment, but the absence of a pair of minor pieces seems to reduce the chances of a decisive result.

I don't understand the c3-c4-c5 moves - surely it leaves the white d pawn backward and firmly blockaded. I think I'd slightly favour black now.

same reaction: it looks there is no point to move to c5... And I would have reserved c4 to a better moment, if ever, for example, Glenda could arrive on h3 and Fou Lenoir on b2. as an example...

Agreed: advancing the pawn all the way to c5 has gifted Black a permanent hold on the powerful posting of the knight at d5, whence it cannot be dislodged. At the same time those centre pawns are immobilised upon squares that the bishop might have found useful. The upside is that e5 and d6 are potentially useful strongpoints for the bishop. Overall Black has a slight edge, but not yet enough to indicate a win.

This doesn't look too promising. How about the imaginative 37...Qh8! attacking e4, but ar=t the same time maintaining the Queen's baleful presence on the h-file. If then 38.Qf2 Qh1+ 39.Qg1 Qf3+ and Black has quite an attack going. White's best seems to centralise the Q (38.Qe4) with hopes of holding off Black's assault.

Oh gosh it looks horrible for the black knight....

Oh, very neat! A classic motif of the bishop's domination of the knight. Can Black save it? Yes: there is a way. In fact, White has to be very careful not to overplay her hand...

I would have been tempted to go after the black knight with my king instead

I don't understand 39. Kf3. If White instead plays Ke3, the king still has the same options of moving to the fourth rank but is also a tempo closer to a2 in case Black takes. That leads to 44. Kxa2 Kxd4 and White has time to defend the c5 pawn.

42. d5 hmmm have you overlooked 42. g5 it basically forces black to move f6 or f5+ something that gives white a little better position

I reckon there isn't much wrong with Jo's choice. Going after the Knight at once won't do: 42.Kd3? a5! and now [a] 43.Bxa5 Nc1+ or [b] 43.Kc4 Nb4 44.Bxb4 axb4 45.Kxb4 f6! 46.Kc4 Ke6 and White is just about in a 'zugzwang'. OK, then, how about 42.g5!? This seems more promising: 42.g5 Ke6 43.Kd3 Kd4!? (43...a5 seems to deserve consideration) 44.Kc2 Kxd4 45.Kb2 Kxc5 46.Kxa2 ... White has the knight, but Black has three pawns for it, albeit scattered and/or backward. My drawing plan for Black would be to commit the bishop to the defence of the g-pawn (46...Kd4 47.Kb3 Ke5 48.Bc3+ Kf5 49.Bf6 ... and then wait for the WK to go after the Q-side pawns, something like 49...Ke6 50.Kc4 Kf5 51.Bd8 Kf4 (51...a5!?) 52.Kc5 a5! It has to be admitted, though, that White's winning chances in this line aren't neglible.