One of your opponents resigned, that was stjohanson.
What happened then is this:
1) You got awarded 1.5 wins against stjohanson (and they got 1.5 losses against you).
2) rickd.canning also got 1.5 wins again stjohanson (and they got 1.5 losses against rickd.canning).
3) There was no rating result recorded between you and rickd.canning.
4) The game was ended right there, no continuation.
4) means only one of your opponent's resigned, not both.
The "tie" in the screenshot is not the best way to describe the situation, agreed.
The general consensus is that multiplayer rated games cannot be meaningfully continued after a player resigns, the bots are not very good per se, and they certainly can't "emulate" the play style of the player that left.
That's why there is a penalty for the resigning player. Note that they are worse off than if they had continued to play and gotten last (in which case the rating would have counted for them 1 loss against you and 1 loss against rickd.canning).
Considering that all 3 of you have no multiplayer rating yet, my assumption is that stjohanson simply didn't know what the effect of the resignation was.
That penalty is there to incentivise players not to do that, because that leads to the problem you found: You would likely have won that game, but now you don't get any result versus rickd.canning.
The problem, of course, is to quantify "who would likely have won". In your game, 3 Provinces and some Duchies and Estates were still in the supply when stjohannson resigned. Theoertically, that could be more than enough to offset your 44 point lead vs. rickd.canning (The game cannot be reloaded unfortunately, so I can't find out for sure). But the board is single gain, so it's not likely that rickd.canning could have gotten most of those cards before the Provinces ran out.
As you can see, all of this requires an actual analyses of the board and the ways to increases (or decrease, say with Swindler) a player's victory point total. That's nothing the computer can evaluate on the spot, at least not with much room for error. And then there are boards that are dominated by VP chips, boards that are likely to end on 3 piles a.s.o.
So the end result of all these problems is that there is no result recorded between the non-resigning players (which is considered the best of all the bad ways to handled this), and resigning is actively discouraged (by having the resigner have 1.5 losses vs. everybody else).