A new hockey week is beginning, so here at OGA we want to look back at some things we have learned to date and what is coming up in the near future.
The Weekend Predictive Experiment
This past Friday through Sunday we attempted to bring you predictions of overall game winners, by period scoring advantages and in some cases, anticipated scorers. Here’s what our work – broadcast on Twitter from @Pelican4 – actually did:
Games: 11.5 Wins; 7 Loses; 62% accuracy. Why 11.5 Wins? We give ourselves a half-game credit for PHX at ANA as a Pick ‘em game that ended in a one goal deficit. You cannot get much closer than that short of a OT/SO decision.
Scoring inside games: 35 Correct within five-minute blocks, to include 1/2 credit for four picks within less than one minute of the predicted block; 103 incorrect; 25.4% accuracy. One of the big hits here was anticipating scorers for a team within a particular five-minute block. We were 6-for-48/12.5% in this regard. As far as players go for simply scoring in that game and not by the five-minute block, we were a more respectable 22-for-46/47.8%.
We believe overall the predictive nature of games should become more accurate as more data from the season becomes available.
Black and Blue Schedule
We are only four games in to the NHL Black and Blue Schedule which has left us a none-too-spectacular 25% accuracy for our four measurement categories. The 22 October COL @ CHI game just about defined our theory, however. In comparison with Game 1 of the pair, the opposite team won, in OT and by one goal. Our lone factor that did not bear out was an increase in PIMs from Game 1 to Game 2 where the overall number of minutes was the same in both contests.
Before the end of the week, Game 2 of TBL versus BUF plays out on 25 October. By theory, BUF should win in OT/SO and/or by one goal with 13 or more total PIMs. Also, the PIT versus NYI and MTL versus BOS series play out, and the first half of DET versus MIN kicks off on 29 October.
The Playoff Qualifying Curve (PQC)
The PQC was engaged for this season on the 22 October as PIT completed their game 10 against NJD. At Game 10, we ranked them as In The Curve – about average, but neither IN (Chasing Stanley) nor OUT (Tee Time) of playoff contention yet. Pittsburgh’s opening schedule was the most difficult in terms of our Fatigue Factors (FF), forcing us to a prediction of 4.5 wins in their first 10 games. As the standings show, and without Malkin and Crosby for a significant or all (respectively) portion of this period, the Penguins overachieved and secured a relative seven wins. The Pens earned 4.5 wins out of six games they were either at a fatigue disadvantage for, or should have outright lost by the model. They are the best example of an overachiever, and are one of the reasons some teams won’t make their predicted number of wins at Game 10.
Another 22 teams will hit their Game 10 this week as the schedule shows below:
26 OCT – VAN
27 OCT – CBJ, DAL, MTL, OTT, PHI and TBL
28 OCT – CAR, COL, EDM and STL
29 OCT – ANA, BOS, BUF, CHI, FLA, LAK, MIN, NSH, PHX, TOR and WPG
How many do we already know a PQC call for? Four right now, but all others need a game or two to clear up the picture. Some need wins, but of the list above, 15 teams will not receive a Tee Time call and two may even receive the early Chasing Stanley moniker.
The final seven teams will hit Game 10 and their first PQC call of the season by 3 NOV.
The Fatigue Factors
We mentioned the FF above in reference to PIT. So you know we are still tracking it to complete our report card soon, nine Eastern and 11 Western teams have met their fatigue minimum range of games projected to win by Game 10. Three each Eastern (PIT, TOR and WSH) and Western (COL, DAL and LAK) have met their projection, and one East (PIT)/two West (DAL and LAK) have met their high range for wins. Three Eastern (BOS, MTL and OTT) and two Western teams (CGY and CBJ) will not make their projected number of wins as of games ending 23 October.
Less than 10 games into the season, what is dominance, really? Maybe advantage is a better word. But here is where team goal for overmatch goals against by official goal-scoring time on the clock. (If there is no mention of your team, there is no clear-cut advantage of three or more goals.)
BOS: 0:00 – 4:49/3rd, 5-1.
BUF: 10:00 – 14:49/2nd, 3-0.
CAR: 5:00 – 9:49/1st, 4-1.
FLA: 15:00 – 20:00/1st, 3-0.
MTL: 10:00 – 14:49/3rd, 4-1.
OTT: 5:00 – 9:49/3rd, 5-1; 15:00 – 20:00/3rd, 9-5.
PHI: 15:00 – 20:00/3rd, 4-1.
PIT: 5:00 – 9:49/1st, 3-0; 0:00 – 4:49/2nd, 5-1.
TBL: 10:00 – 14:49/3rd, 4-0.
TOR: 0:00 – 4:49/2nd, 5-2.
WSH: 15:00 – 20:00/1st, 4-1; 0:00 – 4:49/3rd, 6-1.
CGY: 5:00 – 9:49/1st, 4-1.
CHI: 15:00 – 20:00/1st, 4-0; 15:00 – 20:00/3rd, 3-0.
COL: 10:00 – 14:49/2nd, 3-0.
DET: 10:00 – 14:49/2nd, 3-0.
LAK: 15:00 – 20:00/2nd, 3-0; 15:00 – 20:00/3rd, 3-0.
SJS: 5:00 – 9:49/2nd, 4-0; 15:00 – 20:00/2nd, 4-0.
VAN: 5:00 – 9:49/3rd, 4-1.
And that is our statistical cornucopia for the week. While we have more work to do on predictive, by-game models, we are already seeing some teams showing a variety of advantage over others. We instinctively know and assess the NHL with an eye toward the thought there are only so many Wins and points that can be earned. So when one team – like Pittsburgh – secures more than they rightly should have based on level of scheduling difficulty, other teams projections in the Win column dip. Remember throughout the season, it is all peaks and valleys. You just don’t want your team in a “…Valley so low….”