Pomeroy's is a "predictive" system based largely on Pythagorean projections. He makes several adjustments to the basic points scored vs points allowed quotient. The principles behind his math are covered
here, and he explains some recent revisions and refinements
here.
Without actually crunching the numbers myself, I'm going to assume UMass gets penalized a shit-ton for having so many close games and winning a vast majority of them. This falls within his
"luck" component:
Luck - A measure of the deviation between a team’s actual winning percentage and what one would expect from its game-by-game efficiencies. It’s a Dean Oliver invention. Essentially, a team involved in a lot of close games should not win (or lose) all of them. Those that do will be viewed as lucky (or unlucky).
They've had 13 games with margins of 8 pts or less, and a dozen of them were 6 or less.
W: Neb* 6, Clem* 6, Prov 2, Mia OH 8, St.J 4, SBU 5, @GMU 1, @URI 5, @ GW 6, VCU 5
L: FSU* 5, @Rich 3, @St.J 5
Now let's look at VCU, who have had just 8 such games.
W: @VA 3, LB St* 6, E.Ky 3, @ LaS 8,
L: Gtwn* 4, @St.J 7, @SLU 2, @UMass 5
UMass is 10-3, VCU 4-4. Hence, the Minutemen are "lucky" and not "determined, willful or gutsy cardiac kids", while the Rams have seen the coin flips split 50-50 -- just as KP expects them to.
A major flaw in Pomeroy's system is that it doesn't account for garbage time, when the coach of a team with a significant lead puts in his reserves and sees the margin dwindle. Then there are those final shots scored by the trailing opponents (often their own bench warmers) where they get a dunk or drain a three with a second or two left just to get on the score sheet when the winning team is barely defending, if at all, because the outcome is no longer in question. KP's theory is that such instances should balance out over a season, but it certainly discriminates against better teams -- those who find themselves with double-digit leads most often, and allow them to drop back to single digits. (Likewise, it benefits those that run up the score to build leads of 25+, to the point any garbage time still allows a 15- or 20-point margin. There's no diminishing returns principle, and the teams whose coaches show no mercy don't have a glut of close games where they've won the majority -- so they don't get penalized for it.)
However, "dwindling leads in garbage time" haven't been the issue for UMass. As we've seen, they've had to claw their way back into most of the above games in the final minutes after cold shooting and on-their-heels defense (spawned by the numerous fast breaks off so many missed shots) have put them in early holes.
I don't agree with Pomeroy's system in this regard, but I believe this explains the discrepancy.