You are running into a number of problems with your surface-level analysis here.
I'll address the goals/points after but first of all you are judging PK aptitude on goals against and just leaving it there which is so flawed its hard to take it seriously.
First of all, Marner is a top pairing option on the Leafs which means he faces the best players the other team has to offer. Just to site an example, you do have to admit there is a massive drop off from the 4 forwards Tampa used on their PP1 (Stamkos, Kucherov, Point, Hagel) to their PP2 (Jeannot, Paul, Sheary, Cirelli). A lot also depends on deployment on the PK. With Marner and Kampf being the top 2 guys the Leafs use on the PK, the go-to guy the last 2 years for the 2nd pairing has been Kerfoot (by far the 3rd most PK minutes of the team) and it's striking to see his deployment compared to Marner:
Offensive Zone starts%:
Marner: 1.01
Kerfoot: 8.43
Offensive Zone Faceoffs%
Marner: 1.83
Kerfoot: 11.67
So not only is Marner facing far more elite talent than Kerfoot, he's being buried in the worst possible situations as well.
It's these reasons as to why you'll see this same level of top PKers actually having more goals scored against them most of the time. I mean, it's more or less expected right? The key is to mitigate the most damage possible against the elite players in the league when they have been handed an advantage.
There is also problem with using goals against in general instead of more reliable defensive metrics like shot and scoring chance suppression to evaluate D as goaltending can be a highly unstable thing to look at, especially on a team that has had wildly inconsistent goaltending the last 2 seasons. I mean, poor goaltending can contribute an insane amount to goals/60 when looking at the league as a whole as guys like Barkov (9.2) the wildy underrated Eriksson Ek (9.16) and LAs two top PK guys (Kopitar(9.67), Danault (8.67)) demonstrate. Those guys are not poor PKers, I assure you.
Filtering the last 2 seasons by 250+ minutes (67 players) will give you a far better sample of the players putting up with the tougher usage (although a few 2nd unit guys get through) and looking at those numbers, you have Marner putting up the 2nd most goals/60, 6th in points/60, 3rd most takeaways/60 all while having defensive underlying numbers placing him at 28th for shot suppression, 7th for scoring chance suppression and 6th for fewest high danger scoring chances allowed/60. These numbers would look even better if so many of the 2nd unit guys hadnt snuck in above him in the data.
He's easily one of the better PK guys in the league but the eye test confirms this as well no? A puck hound who is consistently an offensive threat and has the other teams PP on its toes is what I usually see out there.