Abdel Irada: ∞
<<+> File it under "h." <+>>
This was, to me at least, a highly intuitive problem. The key move was the only one I ever looked at, and it came to mind at first glance.
Obviously I'm no GM, but by this point in my chess career, I've seen hundreds of positions like this, many in my own blitz games.
So my thinking went something like this:
"I *could* give check on h1, but that can't be right because the king flees to e2 and escapes. But if I could take control of e2 (and still maintain control of g2), White is staring at mate on the h-file, and not a thing he can do about it."
From there it was all very forthright:
<<+> 21. ...Bf3!
22. gxf3 ... >
Because why not? White has no better option, so he might as well remove the bishop and take his chances.
23. Kg1 ... <>>
Again, what else?
<23. ...Qh8 <+>>
White has run out of resources. It's only a matter of time before his last-ditch defenses, such as 24. Bh2 or Bh4, are annihilated in futility under the guns of the h-file battery.
(Black has a flashier way to finish:
24. Kxh1 ▢, Qh8+
25. Bh2 ... >
There is no hope in <25. Kg1?, Qh3 26. any move, Qg2#>.
26. Rg1 ▢, Rh8>
Not 26. ...Bd6? Rg3! , when White is winning.
<27. Rg7+, Kxg7 >
White has succumbed to horizon effect. He can delay mate for a couple of moves by means of a "brilliant double-rook sacrifice," but he cannot avoid it.
But "flashier" is not synonymous with "most efficient," and this variation is inferior to the quiet 23. ...Qh8, which therefore remains the main line.)
Now we come to perhaps the more theoretically important question: Where did White go wrong?
The culprit may have come as early as 12. Na4. This move looks like a thematic exploitation of the c-file, so I can see why Seirawan was tempted. But it also has the defect of allowing Black's bishop access to e4, which in turn inhibits any move of the f-pawn (which would otherwise create an escape hatch on f2): on either f2-f3 or f2-f4, Black can now take twice on f3 (in the latter case en passant).
I can't say I've established to my own satisfaction that White could not still have saved himself sometime in the following eight moves, but once Black gets in 20. ...Be4, White begins to bust a move Ronda Rousey-style.
From e4, it was inevitable that the bishop would next move to f3, even if White did give up the f-pawn. And that move would seal the fate of his Pallid Majesty, as shown above.