I know that this can break down at the margins (teams that have extremely high/low goal differentials don't have the
extremely high or low win percentages that the pythagorean theorem says they should, but for most normal results, it holds true. If you are averaging 4 goals a game and 2 against, then you're scoring 2/3 of the goals and should win more games than if you are scoring 60% of the goals.
must have been. I was pulling this from the old thread. Sorry about that.
I'm still not 100% clear on this. I'm probably just not wired to get this in the way you're saying it. If there's a point to be made in how it relates to Orr overachieving against bad teams, feel free to go ahead with that and I'll probably understand when it's all filled out... if you still care, which you may not with the voting deadline passed.
Not sold. For example, this gives points for getting a "plus" on the power play (i.e. just being on the ice for your team when they score a power play goal) and the same for a shorthanded point (i.e. having a direct influence on your team scoring a goal while a man down). It seems that one achievement is much more rare and impressive than the other, yet they are rewarded the same. That's why this "stat" looks so random to me. I join C1958 on this one in failing to see its value.
This is disappointing.
I've interacted with you in this forum for nearly ten years now, and I know you are capable of better than this. Contribute to the discussion in a positive way or just don't bother. A person who just wanders into an ongoing discussion with a silly little snipe like this, just looks small.
As far as the subject at hand is concerned, the data that was selected was a one season sample of one team, destined to be full of statistical noise.
From 1968 through 1975, here are the "original 12" franchises based on their goals for and against by period:
| GF1 | GF2 | GF3 | GA1 | GA2 | GA3 | TGF | TGA | Total | TG1 | TG2 | TG3 | G%1 | G%2 | G%3 | diff |
Boston | 816 | 907 | 869 | 546 | 642 | 577 | 2592 | 1765 | 4357 | 1362 | 1549 | 1446 | 31.3 | 35.6 | 33.2 | 4.3 |
Montreal | 763 | 804 | 778 | 491 | 545 | 574 | 2345 | 1610 | 3955 | 1254 | 1349 | 1352 | 31.7 | 34.1 | 34.2 | 2.5 |
NYR | 691 | 758 | 746 | 527 | 571 | 574 | 2195 | 1672 | 3867 | 1218 | 1329 | 1320 | 31.5 | 34.4 | 34.1 | 2.9 |
Chicago | 642 | 771 | 686 | 519 | 531 | 568 | 2099 | 1618 | 3717 | 1161 | 1302 | 1254 | 31.2 | 35.0 | 33.7 | 3.8 |
Philly | 609 | 616 | 588 | 528 | 601 | 562 | 1813 | 1691 | 3504 | 1137 | 1217 | 1150 | 32.4 | 34.7 | 32.8 | 2.3 |
Toronto | 592 | 669 | 662 | 602 | 646 | 624 | 1923 | 1872 | 3795 | 1194 | 1315 | 1286 | 31.5 | 34.7 | 33.9 | 3.2 |
Pitts | 578 | 636 | 618 | 654 | 711 | 666 | 1832 | 2031 | 3863 | 1232 | 1347 | 1284 | 31.9 | 34.9 | 33.2 | 3.0 |
Detroit | 571 | 716 | 692 | 661 | 739 | 744 | 1979 | 2144 | 4123 | 1232 | 1455 | 1436 | 29.9 | 35.3 | 34.8 | 5.4 |
StLouis | 568 | 597 | 579 | 569 | 598 | 581 | 1744 | 1748 | 3492 | 1137 | 1195 | 1160 | 32.6 | 34.2 | 33.2 | 1.7 |
Minnesota | 546 | 565 | 588 | 653 | 689 | 671 | 1699 | 2013 | 3712 | 1199 | 1254 | 1259 | 32.3 | 33.8 | 33.9 | 1.6 |
LA | 497 | 607 | 628 | 644 | 705 | 694 | 1732 | 2043 | 3775 | 1141 | 1312 | 1322 | 30.2 | 34.8 | 35.0 | 4.8 |
Cali | 491 | 546 | 539 | 700 | 810 | 792 | 1576 | 2302 | 3878 | 1191 | 1356 | 1331 | 30.7 | 35.0 | 34.3 | 4.3 |
Totals | 7364 | 8192 | 7973 | 7094 | 7788 | 7627 | 23529 | 22509 | 46038 | 14458 | 15980 | 15600 | 31.4 | 34.7 | 33.9 | 3.3 |
[TBODY]
[/TBODY]
I don't see anything particularly anomalous here. Every team sees about the same amount of goals scored in each period. The idea about slush time is a red herring for a number of reasons:
1. While it's entirely plausible (and has been observed in real life) that teams will play their marginal players more in meaningless times in the game (which tend to be later in the game), the opposite is also true - when it's late and the game is tight, you will rely even more heavily on your best players than you already do. Therefore, over time, there's no reason to believe that any player on the roster will be used more or less in the 3rd than they are in other periods
2. There's no reason to believe that one team has significantly more "slush time" in the long run than another. A very dominant team will be on the positive side of that ledger frequently, and rarely/never on the negative side. A terrible team the opposite team. A team closer to average will sometimes be the windshield, sometimes the bug.
3. There's no data showing that "slush time" has significantly lower or higher scoring per minute - which would throw TOI estimates off for players who play a much greater or much lower than average proportion of their time in slush time. We see times when the goalie gets pulled in a blowout and the ****show just continues. We've also seen teams chase a goalie at 5-0 early in the 2nd... and win 5-0.
4. Every single team follows the exact same trend: most goals in the 2nd, fewest in the 1st, in-between in the 3rd. The differences vary but it's the same principle in every case. Boston is not an outlier.
For the idea of slush time having a big impact on the accuracy of Boston's defensemen's TOI estimates to have any legs, we would have to establish that:
- The total number of GF/GA per minute of slush time was
drastically different from prime time
- They played a huge amount of slush time compared to most/all of the other teams
- They stapled Orr to the bench with the score run up and the game out of reach, and played the likes of Sims, Edestrand and Doak almost exclusively
Then, you
might be able to say that Sims, Edestrand and Doak played an inordinate amount of minutes of garbage time when scoring drastically went up, and Orr played more of his time earlier on, during prime time, when they were running up the score to begin with
The problem is, the first of those things is demonstrably untrue, the second is almost certainly untrue, and the 3rd is impossible to prove with any reliability even if it is true, and if it is, it's counterintuitive, because it presupposes Orr was playing a lot during a time when a lot of goals went in to put the game out of reach (but... then,
more goals were scored in slush time with him on the bench?)
It just doesn't make sense. If you want to go back to what I wrote last night, and really take a look at it, and explain how Orr could have been a 35-40 minute player and have him and everyone else still end up at the same GF/GA figures, be my guest. I don't see any way to get there without making Orr a low-event player - and he wasn't.
Right? He wasn't a low event player.
Right? We at least agree on
that, don't we?
Well sure, in the example you gave, that's a GF% of 50.4% and 50.5%. That's basically the same thing.
15/10 and 25/20, those are two different teams, despite them both being only +5. Those translate to win percentages of .692 and .610.
A team on the margins, let's go crazy and say 6/1, would not actually get a win% of .973. That would be a case of the theorem breaking at the margins.