Let's look at the report.
Current NTRP: 4.0C
Estimated DNTRP: 3.66
Singles Record: -
Doubles Record: 0-7
Sets Won-Lost: 3-14
Games Won-Lost: 59-79
Best Match Result: 4.05 on 3/31/14
Worst Match Result: 3.45 on 3/16/14
Highest Estimated DNTRP: 3.8 on 3/31/14
Lowest Estimated DNTRP: 3.57 on 10/28/13
Singles Average Match Rating: -
Doubles Average Match Rating: 3.72
The 2/1 match was a match tie-break loss played with a lower-half 4.0 against an upper half 4.0 and a 4.0 rated above 4.0 and likely to get bumped up. So they were supposed to lose fairly easily and the close win was better than expected and a good result.
The 3/10 match was played with the same partner against two strong 4.0s and also went to a match tie-break, so same reasoning.
The 3/31 match was played with a 3.5 playing up against two very strong 4.0s and again went to a match tie-break. Getting such a close result with a lower rated partner is a great way to generate a big match rating.
So remember, the NTRP algorithm doesn't really pay attention to win/loss record, instead it looks at the specifics of the ratings of the players involved in the match, the score, and how close the score is to what was expected. Do better than expected, you get a big match rating. Do worse, and the match rating could be lower.