• Schedule Loss

    by Tyler Dellow • April 5, 2013 • Hockey • 7 Comments

    Watching that Oilers game last night, the Oilers looked a little dead. I’m not really a basketball fan but I’ve seen the term “schedule loss” thrown around with basketball – basically, a loss that a team is doomed to suffer because of the schedule. Edmonton played the night before in Calgary and was playing their third game in four nights. The Canucks had two off days before the game. Intuitively, it seems like there’d be an advantage for Vancouver there.

    I went back and looked at all the games between 2005-12 to try and get a sense of how big an advantage it might be. We know that home ice advantage is a thing that’s real: between 2005-12, the home team went 3732-2864-2014 in regulation, a .550 winning percentage. It turns out that there’s a real difference and that being more rested or less rested matters.

    I’m basing rest on the number of days since a team played it’s last game. It’s funny how the home team is 74% more likely to be the more rested team in a game and, indeed, kind of aggravating – while I assume that the NHL would say that it all balances out over the course of the season (and you’d think every team would be screaming at the schedule maker to make sure it would), it’s kind of aggravating that the league systemically tilts the ice in the home team’s favour.

    This is just a high level look at it – I also broke it down more specifically by the amount of rest each team had. The winning percentages are the home team’s winning percentage in games where the road team has had X days rest and the home team had Y days rest. There aren’t a ton of games where one team has had 3 or 4+ days off so there’s a SSS warning on that.

    I’ve highlighted the part of the table that doesn’t really have the SSS problems – 82.7% of NHL games between 2005-12 match a description in that part of the table and it’s exactly what you’d expect – home teams do best when they’ve had two clear days off and the road team played the night before; road teams do best when they’ve had two clear days off and the home team played the night before.

    So, last night was about as tough as it could get for Edmonton. If an average home team had two nights rest and was playing a road team that played the night before, we’d expect them to win in regulation 59% of the time. Vancouver’s probably better than average and the Oilers are probably averageish, maybe a touch worse, so Vancouver probably wins that game in regulation more than 60% of the time.

    The thing with talking about schedule losses, of course, is that they don’t really exist. It wasn’t preordained that Edmonton would lose last night; it was just more likely that they’d lose than in a game against Vancouver where both the Oilers and Canucks had equal rest. It’s more of a schedule tax – a team with a tough schedule in terms of facing rested teams after they play will pay a price in terms of a reduced win expectancy.

    Once you’ve generated these values, it’s easy enough to see who the schedule helps and hurts by virtue of rest – who’s paying schedule tax and who’s collecting it. Oiler fans have spent no small amount of time complaining about the schedule this year – it seems like all 24 of Calgary’s home games are against teams that have played in Edmonton the night before or against the Oilers.

    It turns out that, actually, the amount of schedule tax paid by the most heavily taxed team and collected by the team most favoured by the schedule makers is pretty small. The Maple Leafs have the worst schedule in terms of playing tired and playing rested teams – their schedule would see an average team generate 47.4 points in regulation. The Oilers, of all teams, have the best schedule – their schedule would see the average team pick up 48.4 points in regulation. (It goes without saying but I’m ignoring opponent/team strength in this; this is purely rest and home effects). Calgary’s basically got a bang average schedule.

    I doubt that this is a coincidence – I assume that the NHL has a variety of criteria that their schedules have to meet and I would bet that one of those criteria is that the schedule is roughly equal in terms of playing tired or facing tired teams, because there is a competitive advantage/disadvantage associated with it. I’d guess that they’ve got a spreadsheet or a piece of software that they use to try and balance this.

    The Edmonton/Calgary aspects of this seemed sort of weird to me – the Oilers schedule hasn’t seemed notably friendly this year while Calgary’s has seemed awfully generous, so I looked at games remaining. That would seem to be where the difference lies. I mentioned that the Oilers’ schedule would see an average team generate 48.8 points. Their final eleven games would see an average team generate 11.6 points – most of the Oilers’ schedule advantage comes this month. Calgary, on the other hand, has thirteen games left in which an average team would generate 12.4 points – this implies that their schedule has been easier to date if we know that they end up at average.

    Other notables of interest to Oiler fans: Columbus has the toughest remaining schedule in the NHL. I like what the Blue Jackets did on deadline day but they’re a problem for next year, not this year. Intriguingly, both San Jose (25th) and Detroit (20th) have reasonably tough schedules remaining as well.

    The Oilers are going to need some breaks to get into the playoffs – they’re down to a 27.4% chance of doing so, according to Sports Club Stats but some of the stuff that isn’t quite as obvious does seem to me to be lined up in their favour. I kind of think that Detroit is the target for the Oilers – *mutters expletives about a blown two goal lead at home against Detroit* – and the remaining schedule gives the Oilers an edge.

    Email Tyler Dellow at mc79hockey@gmail.com

    About Tyler Dellow

    7 Responses to Schedule Loss

    1. April 5, 2013 at

      If I’m reading this right, all of your analysis is for regulation play.

      If you include OT (but leave out the shootout), I imagine it won’t change the numbers much, but I’m curious whether the longer games heighten the impact of rest.

    2. Kell Pedersen
      April 5, 2013 at

      You always provide a fantastic lens through which to view the topics from around the water cooler…

      Guys at work were actually talking about difficulty of schedule today in light of why Khabi got the start rather than Dubnyk last night. The question was whether the backend of the back-to-back against the superior team was a good night for resting #40. Considering that the team was likely to be less rested and less likely, at least statistically, to come out on top to begin with, why not run out the optimal lineup in the game you’re more likely to win.

      The fan in all of us wants to see the team commit to running the best lineup every night (re: gimme my money’s worth), but on a night that the team likely would be outscored by two goals, why not run out the backup; if he gives up that one extra goal per hundred (http://www.mc79hockey.com/?p=5583) during the game, well then, you lose by three instead. As we saw last night.

    3. Ryan V
      April 5, 2013 at

      Simple idea, but those are great numbers. It’s always exciting (for me, at least) when the numbers are totally unsurprising and make complete sense.

      Fans of Western teams tend to complain about the length of travel and time zone changes. I wonder what that table would look like if you ran it only for Western teams playing on the road vs Eastern teams, and vice versa. Do rest and road games matter more in one conference than the other?

    4. Mike
      April 5, 2013 at

      @Kell – absolutely agree. Nothing frustrates me more than when the team runs the backup in the “easier” game and manages to lose both.

    5. Bruce McCurdy
      April 6, 2013 at

      Good stuff, Tyler. Just read this now after hearing your (excellent!) spot on Lowetide’s show today.

      I’ve long been (over-)sensitive to scheduling/rest issues but don’t have the mad skills to crunch big data like this. Nice work. I guess there’s more to home-ice advantage than line-matching, eh.

      I had just written at the Cult of Hockey this morning about Oilers being the less-rested team in recent games, using a slightly different method. Days off since last game -> if same day then previous game -> repeat. So by that method the Kings with their Tue-Thu-Sat sked have a slight edge on the Oil with the Wed-Thu-Sat slog, even though both teams have had one day off since their last game. Anyway, the current game will be the ninth in a row where the Oil had less (short-term) rest than the other guy. Have to think that’s significant especially in a compressed season.

      I look forward to that turning around down the stretch as alluded to above, haven’t really looked ahead that far but something’s gotta give at some point.

    6. Tom Benjamin
      April 6, 2013 at

      Great stuff, Tyler. When I was looking at this stuff in the nineties, I defined tired team as playing three in four nights ending with the back to back. The winning percentage for the rested team was really high – about 75%.

      The Maple Leafs have the worst schedule in terms of playing tired and playing rested teams – their schedule would see an average team generate 47.4 points in regulation. The Oilers, of all teams, have the best schedule – their schedule would see the average team pick up 48.4 points in regulation. (It goes without saying but I’m ignoring opponent/team strength in this; this is purely rest and home effects). Calgary’s basically got a bang average schedule.

      I doubt that this is a coincidence – I assume that the NHL has a variety of criteria that their schedules have to meet and I would bet that one of those criteria is that the schedule is roughly equal in terms of playing tired or facing tired teams, because there is a competitive advantage/disadvantage associated with it. I’d guess that they’ve got a spreadsheet or a piece of software that they use to try and balance this.

      The NHL sure wasn’t trying to balance anything in the mid-nineties. True Western teams were much more likely to play tired against rested teams. It wasn’t even uncommon for Western teams to come off a long road trip and play tired against a rested team at home. I know there has been a lot of whining about this in subsequent years and the relocations have generally helped reduce the geographic disparities so it may be different now. The league is still geographically skewed to the East so I think this year’s results may very well be a coincidence, one that’s driven mostly by the lack of interleague play.

      What did the Oilers pay (or collect) in the schedule tax over the seven years before this year? Who paid the most? The least?

    7. Nathan Daniel
      April 7, 2013 at

      I’ve often wondered why the home ice advantage seemed to be less prominent in the playoffs.

      It seems to me that in the playoffs, the teams would be rested an equal amount, except for the first game of each round, where there’s a good chance that one team has more rest, though not always the home team.

      As such, one of the ‘baked-in’ reasons for home ice success is removed.

    Leave a Reply

    Your email address will not be published. Required fields are marked *