That one bakes in a factor for how the team did when a player was not on the ice.
It makes guys like Bourque and Jagr look really good because their teams were had little depth during their prime years and they carried their teams.
So it does reflect well on guys like that but it doesn't necessarily tell you who was "better" between a couple of players.
Lidstrom doesn't get as good a boost because his teams had depth. Similarly a guy like Larry Robinson who has the highest raw +/- gets knocked down a notch for the same reason. The players he is being compared against were stronger so it is more difficult to separate oneself.
I agree with this. That kind of "adjusted +/-" has some benefits, but it also have some disadvantages, like the ones you mention. It should preferably be looked at within context (like "all" stats).
A problem with ES+/- is that it basically makes all defencemen as good as all forwards (on the team), and all wingers as good as all centers.
The average ES+/- of the left wings will be the same as the centers, no matter if the centers are Sakic/Forsberg and the left wings themselves in reality all are below-average players.
The "average" ES+/- of the defencemen will be the same as the sum of the forwards, no matter the forwards are weak while the defencemen are the best ones in the league.
This is a basic fact that is important to keep remember.
The rOn/rOff thus mostly tells us how good a player performed compared to the other teammates playing on the same position as him. In other words, the player(s) that step onto the ice to replace him when he goes for a rest.
If Bourque had poor teammates, his rOn/rOff would be compared to that poor defenceman replacing him. It could be the 2nd best defenceman on the team, but it could also be the 3rd, 4th or even 5th.
It is thus no wonder that Bourque would have a significantly better rOn/rOff than the other defencemen on the team.
For example, Borje Salming was in a similar situation in Toronto. In his case, I think he also often played with the clearly best forward line on the team.
A player "hurt" by the stat might be the 3rd and 4th centers on Colorado, who even if being above-average players in reality, would have to see themselves compared to Sakic and Forsberg (who also often had better wingers).
Along with the above also comes the other often mentioned weaknesses of +/-.
I wrote about this in 2011. The idea might behind the rOn/rOff might seem good, but in reality it doesn't really help as much as one would think. (If one wants to adjust +/-, a better starting approach would for example be to try to adjust the "raw" +/- for goalie influence.) I think points scored and icetime are more "reliable" stats. Defensive performance is unfortunately difficult to measure.