First off, i have to apologize for not reading your post thoroughly: i missed the part where you said the regular season does not matter.
Second, i want to ask you where you got those numbers from? Did you go back to watching all those games or are the stats on the internet somewhere, because i can't find the +/- stats for the Lakers series specifically. But looking at the 2004 playoff stats (
http://www.82games.com/034PSAS.HTM ), Devin, although improving on his regular season ratings, still has a negative +/- rating for the whole playoffs combined. The team was +0.4 (+1 total) with him on the court and +4.2 (+30 total) while Devin was on the bench, combining for a -3.8 (-29 total) Rolland Rating. This could be explainable if Devin had a really bad series against Memphis, but he only played a combined 11 minutes in 3 games Pop put him on the floor.
So if the team was +19 when Devin was in the game (119 minutes) during those 6 games against the Lakers and -44 when he wasn't (169 minutes), the team per 48 minutes would be +46 (288/119 x 19) with Devin and -75 (288/169 x -44) without him, combining for a total rating of +121 (not +149).
Now if that was truly the case and Devin really had a +19 and -44 net total (+63 combined), then Devin, to end up with a total of -29 point rating (
http://www.82games.com/P3SAS4D.HTM ) would have to help the Spurs go for -92 points in the whole 11 minutes that he played against the Grizzlies. How that is possible is beyond me, but from the stats i can see, in two regular seasons and two playoffs with the Spurs, Devin has yet to produce a positive Rolland rating.