Not true. There is a correlation, but it's not a direct 1:1 correlation. Teams can over- or under-perform their goal differential.
For instance, look at 2012. The Isles had a -52 GD And finished ahead of Montreal which was -14. The Isles finished higher despite having a GD that was 38 goals worse - nearly half a goal per game worse.
That same year, Florida had a -24 GD and won their division. They finished two places higher in the league standings than Ottawa, which was +9 (33 goal difference).
My point is, having a worse GD than team X is no guarantee that we will finish behind team X in the standings.
According to sportsclubstats, our odds of finishing 30th overall are down to 46%.
Panic button time.
And stop telling me about goal differential and Corsi and possession stats. I want to see it in the standings. That's what matters. They are not going to be assigning lottery odds based on goal differential.
You're basically arguing that imperfect data is useless because sometimes it is wrong. There is variance/spread in any correlation. The point isn't to see a few instances where the data doesn't fit and declare the correlation unreliable, the point is to look at it statistically. "No guarentee" is a pretty worthless argument. Nothing is ever guarenteed, and the only people who never seem to realize that are the same people always using the "No guarentee" argument. Your Florida/Ottowa example isn't even a particularly good one, as Florida led the league in OT/SO losses that year, and if you take 8 of those OT losses and make them regulation losses to match Ottawa (keeps their same goal diff), then they drop in points to 86 and 20th place, right around where they'd be expected to be based on their goal differential. But of course you should expect the correlation to not be as good the closer to even goal differential you get, when quirks like who gets more loser points comes into play. But that is irrelevant to this discussion, as we're talking about whether goal differential can predict whether the sabres will finish last.
Since the Sabres existed, 35 of the last 43 seasons the team with the worst goal differential finished in last place, or 81.4% of the time. In the last 20 seasons, the "success rate" is 17/20, or 85%. In the last 10 seasons, only 1 time has the team with the worst goal differential not finished in last place (90% "success"), and in each of the past 5 seasons the team with the worst goal differential finished last (100%).
With this data, I can also tell you that in every case the team with the worst goal differential did not finish last, the team that did finish last was within 0.2375 goal differential/game, or within ~19-20 goals in an 82 game season. Right now after 21 games, the sabres are 0.4286 goal diff/game worse than the next closest team in goal diff - almost twice the margin the Washington Capitals had when they got beat out for last place by the Minnesota North Stars in 1978.
What this means is that if the sabres finish worst in goal differential by more than 20 goals, which they're well on their way to doing, then historically they are practically guarenteed to finish last in the standings. Before I'd even worry a little about the sabres not finishing in last, they'd have to get a lot closer in goal differential to the next closest team (within about 5 goals at this point of the season). If the sabres end up -133 and the Oilers end up -98 like their current goal differentials project to, it is very unlikely the oilers would be tied or behind the sabres. A team that goes -70 the rest of the year is not going to lose as much as a team that goes -100 the rest of the year. Also take into account that the sabres are on a 3 game winning streak while the oilers are on a 4 game regulation losing streak and still they're just tied in points... well at some point this season, the streaks will go back the other way.