I tried to figure this out, but I'm also at a loss. I did find a short review of the book, and something was said about "winning with the pawns", but the explanation was short, murky, and cryptic.
I guess the takeaway for me is to make a mental note never to buy this book. I mean, if the author can't clearly define his ideas, then he deserves to have people ignore his books.
I'm intrigued. It sounds both lurid and stupid at the same time.
Here's the quote from the review:
The term "pawns and goat pegs" begs an explanation. To do so Mc Donald gives a quote from a Turkish manuscript written in Constantinople in 1501.
"The fourth arrangement is called al-fazz or gechi gazighi (goat peg). It is so-called because he who plays it wins with the pawns. They are like a peg in his opponent's clothes, and the opponent is like a man with his hands bound."
Great. I now have a Turkish quote from long ago, but I'm no closer to understanding what it actually is.
There are many maneuvers in pawn jiu jitsu. But I've never figured a "goat peg" and can't tell from the examples.
Thanks for the responses. I thought somebody here would have known about this topic already, so I was surprised. Here is the intro to that chapter:
----------
(p. 150)
5. Pawns and goat pegs
'The fourth arrangement is called
watad al-fazz or gechi gazighi (goat
peg). It is so called because he who
plays it wins with the pawns. They
are like a peg in his opponent's
clothes, and the opponent is like a
man with his hands bound'.
From a Turkish manuscript
written in Constantinople, 1501.
As well as the 'goat peg', opening
strategy in the Arabian form
of chess also included the torrent
pawn: a pawn advanced like a
battering ram to splinter the
opponent's pawn structure. The
game Grischuk-Kamsky, given in
the Ferocious Files chapter,
witnessed both a goat peg on d5,
which restrained the black knight,
and a torrent pawn that charged
down the h-file.
In the present chapter we shall
look at more examples that
demonstrate the spirit of the ancient
game is alive and kicking in modern
chess strategy.
----------
In the game annontations the author also identifies a "torrent pawn," so it sounds like he's mixing two concepts in the same chapter. (I'll update my posted games to distinguish between these--sorry.) Also, it sounds like the identified pawns were sometimes only *threatening* to become goat pegs, especially in Game 24, where the annotation says at one point:
(p. 151)
16 c5!
The torrent pawn mentioned in
the introduction to the chapter. Now
(p. 152)
the white pawns are cramping the
black queenside pieces and there is
the latent threat of c5-c6 (when the
pawn becomes a goat peg!).
Seems quite clear to me. A pawn advanced like a battering ram to splinter the opponent's structure.
For instance, in Kramnik vs. Topalov, 17.f5 is a goat peg.
I think it's when a pawn ventures into the enemy lines and becomes a nuisance that he can't get rid of easily or without doubling up pawns etc.
Just a single pawn blocking his pawns, covering important squares etc. I think the most useful example is when you advance a pawn to threaten an enemy pawn (to open a file or something) and he doesn't exchange, instead he advances his pawn and sticks it in front of another pawn, usually in the 3rd rank from his enemy's perspective.
Come on, we've all had that happen, right?
Yes, now that I realized he's mixing two concepts (that must be why the chapter title is "Pawns and goat pegs" and not just "Goat pegs", but in that case why didn't he entitle it "Torrent pawns and goat pegs"?), it's becoming clearer that the pawn sacrifices (yes, we've all had that happen, but I've just heard them called "pawn attacks" or "pawn storms", so why did he use new terminology?) are the torrent pawns and the cramping pawns are goat pegs. So why didn't he just call them "cramping pawns"? And why didn't he explicitly mention in the intro that one can turn into the other instead of leaving that critical knowledge to a single annotation in a single game? I wouldn't recommend this book for that reason, although the sample games are probably good.
I think if the pawn can't be taken by the enemy and just sits there being annoying, it's a goat peg. If you just rush pawns aggressively forward for sacs/exchanges, to double up his pawns, drive his minor pieces away, open files, trap or expose the king etc, then it's just a pawn storm/torrent. Not a peg.
A goat peg from what I understand is like a spike you stick in the ground and you tie the goat to it. So it means that it's restricting movement. So a pawn-peg should not be disposed of easily, it should remain lodged in front or inside the enemy formation.
That's how I think about it anyways, today is the first time I've heard of this term, heh 
I knew exactly what you were talking about from the title. I loved the McDonald books. I seem to recall he did explain what they were in the book though.
21 f6! is the goat peg in the game you said you couldn't see it. A goat peg is a pawn that restricts an enemy piece and keeps it from entering the game easily, and in this case shuts the bishop down altogether.
21 f6! is the goat peg in the game you said you couldn't see it.
Thanks. I updated my diagram and its focus position in that posted game above to reflect your discovery. Actually I also later guessed that that was the goat peg, but I forgot to update that diagram when I updated the other diagrams, but I'll give you the credit for finding it.
P.S.--By coincidence I just came across this great Alekhine game in the past hour that used a goat peg very similar to the one in Game 25 posted above:
Goat peg. Bad bishop. Tall pawn.
I still say this book's presentation was lousy. There should have been two separate chapters, one on each type of pawn (goat peg, torrent), or at least a subdivided chapter, each with its own title that clearly described the topic. Even though now I understand what the author meant, I'm still irritated because each type of those pawns encompasses a span of several subtypes of pawns. For example, I thought a goat pawn was a blockading pawn that kept a pawn from advancing to free the piece(s) behind it--totally *prevented* a pawn advance--but he's also including pawns that only *discourage* a pawn advance, and probably other subtypes as well.
I've seen this book recommended a couple times on this site...
McDonald, Neil. 2006. The Art of Planning in Chess: Move by Move. London, United Kingdom: Batsford.
...so I took a look and was intrigued by a term I hadn't heard before: "goat peg." However, I was really irritated that the book didn't outright explain what a goat peg was, but instead the book just provided several annotated games. Here are those games.
(The goat peg was found here by latvianlover:)
Does somebody want to venture a definition of a "goat peg" and/or find the missing goat pegs? The best I can determine so far is that a goat peg is the same as a "cramping pawn."