Interesting comments. They got me to haul Lewis off the shelf and have a quick glance through. I'm posting one diagram here, a spiderweb plot comparing Guiness, Beamish and Murphy's, neither of which I like nearly as much a Guiness - de gustibus non est disputandem I suppose. I'm not quite in agreement with the fellow that wrote "The beer [from Cork] is by far the worst I have ever tasted." but I'll take a Guiness over a Murphy's or Beamish any time it's available. But that's my personal taste. I'll come back to that.
Anyway there are substantial differences in the hoppy, sweet and fruity/estery axes none of which you mentioned as being responsible for the supposedly poor quality of Guiness. Astringent/phenolic variation is much less than these as is roasted.
I also found the acidity comment interesting as while it is true that Guiness does contain somewhat more titratable acidity than the other 2 Murphy's pH is lower at 3.97. And while on the subject of pH (to which the palate responds - it doesn't care about the buffering behind it)
I'm also attaching the cumulative pH distribution of the 21 stouts Lewis studied. The data show the median pH to be 4.25 and the mean 4.02. Stouts are sour beers. Note that the pH of Guiness at 4 puts it in the bottom decile whereas Beamish at 4.3 is in the 40%tile and Murphy's is off the chart to the left at 3.88.
As to the alkalinity issue. Lewis brewed 9 experimental beers in which he varied the amounts of roast barley and used 3 different waters: DI, DI + CaSO4 and simulated Dublin. Rather than paraphrase I'll quote:
"We found that when additional roasted material was added it increased the roasted and burnt flavor and aroma characters - not a very surprising result. Furthermore, the more alkaline water consistently increased the total sensory bitterness of the stouts, though as far as we could tell it did not much change the sensory quality of that bitterness. These two effects were quite marked, but no other consistent sensory effect of brewing water showed up in these trials."
But that's sensory. The pH's of the beers made with more alkaline water were higher (that's a note to Kai if he is reading this).
So higher alkalinity water doesn't make a better stout. It makes one that Martin prefers. This is classic confirmation (or cognigtive) bias. It is hard as hell to fight that and win (I can't). Guiness isn't bad beer or a bad example of a stout. It is typical (look at the other spider diagrams and the PC plots in Lewis's book if you have it. Nor, though he calls Guiness "the quintessential stout" does he say it is the best stout. It is a typical stout. I prefer it to Murphys or Beamish and I prefer the qualities that Martin doesn't like - in particular the dry tartness with the hint of coffee and of course the head. But I won't tell anyone that he has to like it or that Murphy's and Beamish aren't good beers. They can't be as they enjoy pretty wide markets.