WS measures the supposed impact of a player on the offensive and defensive side to estimate how many games a player was responsible for winning the teams games. Sure it's not an exact science, and there are exceptions, but overall, it provides a very good guidance to which player is better. A player, even a superstar, requires a certain set of teammates to maximize your potential, so naturally quality teammates does help with your WS, but it doesn't mean that having the best teammates will lead to better winshares. Garnett had way better teammates with the Celtics, but the Celtics didn't center their entire offense and defense around him (coupled with the fact that Garnett was older then as well) so his WS was less than those of his Wolves day.
A player with a high WS on a winning team usually translates to a player with the ability to have a winning team constructed around him. So a consistently high WS speaks to a player who is relatively versatile and easy to build around, while a player with wildly fluctuating WS are those who are system based.
Taking Garnett, Dirk, and Duncan as comparisons. During their primes, they have approximate WS ranges of:
Garnett: 11.6 to 18.3, with three fantastic seasons while they had Cassell and Sprewell.
Dirk: 11.5 to 17.7, but he had consistent runs of 13 to 15 in there with many different types of teammates.
Duncan: 10.1 to 17.8, and again, with consistent runs of 13 ws, with no sudden ups and downs. Also note that Duncan plays generally less minutes as part of the Spurs strategy
I tried to put in Shaq, but he didn't play full strength or were injured quite often, so it's really tough to put that in perspective.