• Zuerst kamen sie für die Flammen

    by  • March 19, 2007 • Uncategorized • 7 Comments

    I hope the Flames miss the playoffs. I’ll take it but I didn’t want to see it happen this way.

    Calgary is in grave danger of missing the playoffs, almost entirely because of the ridiculous manner in which the NHL hands out OTL/SOW points. The Flames are, by Andy Dolphin’s reckoning, the best team in the NW and by Sagarin’s reckoning, they’re number two. I suspect that Sagarin is using the GF/GA numbers published by the NHL, which include SOGF and SOGA, and would accordingly bugger up the system.

    I’ve put together a chart summarizing the OT/SO information for this season. The table is self-explanatory I think, with the exception of the last two columns. XP means extra points. I treat SOW and OTW as extra points, which would not be traditionally available. While it’s obvious that the SOW is the extra point as compared to hockey tradition, it’s less obvious that the OTW should be treated as the extra point. My thinking in treating it, as opposed to the OTL point as the extra point, is that, like the shootout, OT hockey is not the same as regular hockey. Not only is it 4 on 4, but it’s 4 on 4 with a drastically different set of incentives than those in place in a traditional hockey game. There’s no risk of the OTL point being lost – there’s only an OTW point to be gained. Accordingly, I think it’s properly viewed as the extra point. This is a switch for me – I didn’t always view it this way. XP/OTG is just a measure of how effective teams have been at extracting the available extra points from their OT/SO games.

    You can see how Calgary has suffered. The Flames have three points out of their OT/SO games compared to sixteen for Vancouver and seventeen for Minnesota. With Calgary eight and seven points back of those teams respectively, their results in the OT games are the reason that they’re in eighth instead of sitting in the top three. The three point OTW/SOW edge that the Avs have on them is making the race a lot closer as well. I’m no rocket surgeon but it seems insane to me that the Flames might miss the playoffs, a test of a specific form of hockey skill, on the basis of subsets of hockey skill that aren’t used in the playoffs.There’s more to it than just the absurdity of determining who makes the playoffs on the basis of tests of hockey skill that aren’t applied there. There’s equality of opportunity as well. About half of the edge enjoyed by the Canucks and Wild in OT results is pure opportunity. They’ve had more games go to OT than have the Flames. It strikes me as more than a little unfair that the Canucks and Wild should be rewarded for their inability to put a team away by playing the rules that God intended with an opportunity for an extra point in situation.

    We were discussing this topic in the Alberta Baseball Confederacy draft room last night – I suspect that’s what led to Matt’s post here – and Andy, once he’d been convinced that the SO/OT points impacted the thing, disagreed with the premise that luck plays a huge role in determining whether or not teams get the opportunity to rack up bonus points or not. He doesn’t think that luck matters any more than it does in terms of any other hockey result. I vehemently disagree with this position.

    I wanted a fairly reasonable dataset to look at, with consistent rules in force, something that’s disappointingly difficult to come across in the NHL. I chose to look at games played between 1999-00 and 2003-04, when the OTL was in force but before the shootout. Fans of Bill James and baseball research will be familiar with the idea that for something to be a skill, it has to be repeateable. If you can’t show that something is repeatable, it suggests that it’s not a skill but rather, that it’s due to outside factors.

    I looked at team results in a number of different categories from 1999-00 to 2003-04. The categories I chose were regulation wins, regulation losses, regulation ties and OT point percentage. I’ve put together a table with each team’s results in a given category in one season along with their results in that category for the next year. For instance, the Oilers won 28 games in regulation in 1999-00 and 34 games in regulation in 2000-01. The next line down will have their 34 wins in 2000-01 compared to their 35 regulation wins in 2001-02. Repeat for all seasons and all NHL teams in that time period. Then, I pulled correlations between the results. My idea is that a greater correlation in one category will indicate that the result is more of an “ability” than a different category with a lower correlation.

    So what are the results? Regulation wins and losses are far more repeatable – they had coefficients of correlation at 0.55 and 0.64 respectively. Regulation ties came in 0.16 and OT points percentage at 0.14. I can’t say from this that regulation ties and then mining points in OT is necessarily luck based but I do think it supports my view that it is. Criticisms are of course readily available – a team changes from year to the next and maybe the sample size is too small to detect the “skill” but I’m still inclined to think that it’s largely influenced by luck. I’m also inclined to think that the Flames – like the Oilers, Hurricanes and Senators – have been caught in a bit of a perfect storm this year. They’ve been crappy in OT/SO play and haven’t had many opportunities.

    I wrote something a long time ago arguing that the NHL would probably like this because it creates the illusion of more parity in the standings. Last year’s Stanley Cup Champs might miss the playoffs! If the Flames make the playoffs and make a run – It’s the eighth seed making a run! It’s the NHL – anything can happen! It’s at the expense of proficiency as a determining factor in which teams make the playoffs and which teams reap the benefits of home ice advantage though. It’s bad enough that the NHL has such a distorted schedule, with teams competing for playoff berths against teams playing dramatically different schedules. The OT/SO results just push the outcomes even further from skill towards chance. It’s a shame that there’s nobody within the game who will argue in favour of competitive integrity as a value that needs to be protected.

    About

    7 Responses to Zuerst kamen sie für die Flammen

    1. lowetide
      March 20, 2007 at

      I’m hoping the Flames make the playoffs and have a deep run. In fact, if they win the Cup it would be a good thing imo. The EIG would have sweat running down the crack of their ass all summer if Iginla raises the Silver Cup.

    2. March 20, 2007 at

      Interesting take. Somewhat refreshing as compared to the glum cloud that hangs over the majority of Flames fans currently.

    3. ChrissyT
      March 20, 2007 at

      I think last year being the first year of the OT/SO format, teams could rightly claim to not having grasped it yet, but this is year two, and that excuse is getting old fast. To me, Calgary has a good team on paper for OT or SO, and getting the XP/OTG. They’re just not getting it done, Vancouver, Minny, and Colorado are.

      Colorado could make the same argument as Calgary (ie. “well Van + Minny play more OT than us, so…”), they’ve played 1 less OT/SO games than the Flames.

      What was Calgary’s record in OT/SO last season? I’m thinking they weren’t this horrible (.230) at getting the extra point. Vancouver (11-3) and Minnesota (7-1) both have outstanding OT records compared to Calgary (1-5). I also find it strange that the Flames are worse in OT than in the SO (2-5).

      Let’s suppose then for a moment that Calgary and Colorado had both played 10 more OT games. Add on 10 points just for being there, and given Colorado is at .500 right now, they’d probably get 5 extra points. Calgary would be lucky to get 3 extra points, so Colorado would only be 2 points behind Calgary for 8th spot!

      I say give credit to the Avs for still hanging in the race. The NW was a pretty close division all year, until the Oilers stopped floundering and finally drowned. IMO, if the Avs take that 8th spot from the Flames, good on them. The real problem with the Flames is, and has been, their road play. They’ve had ample opportunity to put the Avs out of their misery in the past week or so, they just don’t show that killer instinct. Didn’t in last years playoffs, nor in the post-lockout SCF.

      This year, the Flames haven’t played well enough this year to earn the division title, IMO. I still think Minny will win it. Minny is a totally different team with Gaborik, they play close to .700 hockey with him in their line-up. Luongo’s been the difference in Vancouver, and Colorado is still paying Jose close to 5M to back-up Budaj (ouch!).

      I thought Calgary was a strong team on paper without trading for Conroy, Stuart, and Primeau, and it must be frustrating for Flames fans to be in the position they are in right now, but if Calgary could even get .500 XP/OTG they wouldn’t be in this mess right now, cause they’d most likely be 3-4 more points up on Colorado.

      I think the Flames would have more of a case for arguing the schedule is detrimental to them. IMO, the NW is the toughest division in hockey, hands down. It certainly can be argued that the NW has 2 of the top 3 goalies in the league, and Backstrom is right up there with Kipper and Luongo this year in Sv%. Aside from Edmonton, all the other NW teams have excellent GF/GA differentials. Nucks (+17), Minny (+30), Calgary (+31), and Colorado (+14). The Oilers are an ugly -36, but keep in mind that in this 11 game winless streak the Oilers are -29. Ugh.

      GF/GA differential by division:

      N West: 1056GF, 1000GA, +56
      N East: 1161GF, 1132GA, +29
      Pacific: 1056GF, 1045GA, +11
      Central: 1021GF, 1039GA, -18
      Atlantic: 1056GF, 1075GA, -19
      S East: 1099GF, 1158GA, -59

      Sorry for getting so far off-topic.

    4. antro
      March 20, 2007 at

      in order to make the argument about the low coefficient of correlation showing “luck”, you’d have to research which teams were winning against which teams (ie, are they strong or weak teams). presumably the strongest teams were going to overtime less often, and even then, one would want to know against who. that is, you’d somehow need an index of relative strength of both teams which went to overtime, and then figure out if that explains the winner more often than not. (given that teams get on hot streaks, i’m even tempted to say you’d need a time-sensitive relative measure that assesses the teams at the date of play if you really wanted to improve the explanatory power of your model). since a majority of games are still decided in regulation, it’s not surprising that you are getting a higher coefficient of correlation for regulation wins. the strongest teams are more likely to be the winningest. by not factoring in some kind of relative strength for the subset of teams that go to overtime, your “luck” coefficient is spurious.

      one idea might be to use the goal differential of the team at time of overtime game, since using the position in the standing would be circular.

    5. oilswell
      March 25, 2007 at

      God intended even strength 5 skaters only, this PP foolishness rewards a limited skill set of less than a third of the team.

      If every game was decided by shootout you’d have a lot more solid data to work with.

    6. mc79hockey
      March 25, 2007 at

      They’re both responses to specific evils, I suppose oilswell. I tend to think that the evil of a tie is significantly less than the evil of an uncalled legit penalty.

    7. July 29, 2007 at

      This is exactly what I expected to find out after reading the title . Thanks for informative article

    Leave a Reply

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