NotProkofievian
Registered User
- Nov 29, 2011
- 24,490
- 24,640
We needed a better goal differential. Believe it or not some teams have won playoffs series scoring fewer goals than we did against the NYR. Obviously the bigger and easier improvement to make is by boosting offense but if Emelin/Markov/Beaulieu don't make mistakes left and right... perhaps we win against NYR.
Going to go ahead and disagree with you here. We need more goals. First, it's very difficult to give up fewer goals against than we did this season, or than we did in the series against the rangers against playoff opposition with any kind of repeatability. Given that we had so much trouble scoring against quality opposition last year, and that we let in so few goals, from a ''principle of limiting returns'' perspective the move seems obvious: strengthen your weakness.
From a less obvious perspective, let's assume that a team's GF against any other team is going to converge to the average against a representative sample: say, against playoff teams, or teams below a certain GAA. Then a team's GF/GA is going to converge towards GF1/GF2, where GF1 is their own GF against representative teams, and GF2 is that of their opponents. For the same difference between GF1 and GF2, the team with the smaller GF will win a greater percentage of games and series the larger their GF is.
This is because of integer effects: you can't score fractional number of goals. What happens is that you let in/score a certain number of goals a different proportion of the time. Try as you may, your team's GAA will not be too much lower than 2 the longer your sample runs. That directly affects how many times your team lets in a certain number of goals: say 2. If your team lets in 2 goals, that's a huge deal because your team now needs to score 3 to win, which is actually way less likely of an even than scoring 2 is. It doesn't really matter if your GAA is 1.86 or 1.95, you're going to have to face the reality of scoring 3 to win roughly the same amount of the time. Thus, your best bet is to increase the likelihood of that happening. For us, it just wasn't that likely.