I posted the Moulson/Ennis splits with Girgs vs Larsson a few pages back. They were better offensively with Larsson, and SIGNIFICANTLY better defensively as well.
Thanks for the link to the stats by the way. Unfortunately, it didn't really have what I was looking for. I was looking for the game by game raw data to look at trends over time and strength of opponents and what not. I could probably spend a few hours manually going through the data on NHL.com but I don't really have that time now.
My suspicion is that the easier schedule after the trade deadline is primarily what makes Ennis and Moulson stats with Larsson look better than with Girgs. It's not a knock on Larsson, its just a smaller sample size where the team was playing better than the average of the previous ~60 games, which had ups and downs.
One of the things that would be influencing Girgs stats, but not Larsson, is the first 20-ish games where the team played its worst hockey collectively against the toughest competition. I remember Girgs and Ennis still playing well relatively during that stretch, but it would make sense that possession stats would suffer (see: Myers, Tyler). I also remember Girgs had a stretch of 10-20 games where he got cold for a bit, which would influence his stats over a still small sample size of ~60 games he played this season, but going forward may just be a blip or growing pains. I would expect Larsson to go through similar struggles over larger sample sizes against tougher opponents as well. There's also how the other roster players moved around to consider. Girgs I'm guessing was on the ice a lot with Myers and Gorges, and Larsson probably had Risto and Bogo a lot. I'm not sure how that would affect things, but my guess is the latter ended up with better possession stats.
One of the things I would be interested in seeing is something like comparing the 10-15 best "with girgs" games to the end of season "with Larsson" games, rather than an average. I just feel like the sample sizes are maybe too small and too different to make a comparison of what are essentially season average stats. The games Larsson played as top C probably aren't necessarily equivalent to the 60ish games girgs played as top C.
I feel like I should put a full disclaimer here that this is many just me spitballing going off my imperfect recollections of watching the games this year, so I don't need anyone jumping down my throat if my suspicions are incorrect. They're just the things I would be looking to confirm or disprove if I had time to look into the stats a little more. I'm not trying to argue one player over the other or anything like that, only that I'm not convinced that stats you showed us prove that Ennis and Moulson were absolutely more productive with Larsson - only that they were more productive over the last 15-20 games on average than they were the previous ~60, on average. I was certainly impressed with Larsson and I remember thinking at one point while watching the games that Larsson was providing about the same impact on the top line as Girgs was.