One of the things that I think would be the most interesting part of any general manager’s job is the struggle to understand where your team is now, in terms of performance and the challenge of getting your team to where you want it to be - presumably, a lengthy spell amongst the NHL elite.
This topic came to my mind a week or so ago, as I had Stauffer’s show on for some background noise while doing some work. They were talking about the Oiler and, if I recall correctly, Stauffer was expressing the opinion that an Oilers playoff miss next year should jeopardize the careers of Kevin Lowe and Craig MacTavish. It might have had something to do with Darryl Katz; I don’t quite recall. Anyway, it was interesting to me because I kind of think that the playoffs are further away than they appear. You can’t really get a solid handle on that though without having some sort of a handle on what really happened this year.
There’s a cool passage in Moneyball where Michael Lewis talks about run differential in baseball with Paul DePodesta, then the assistant GM of the A’s, now living in his mother’s basement.
Before the 2002 season, Paul DePodesta had reduced the coming six months to a math problem. He judged how many wins it would take to make the play-offs: 95. He then calculated how many more runs the Oakland A’s would need to score than they allowed to win 95 games: 135…Then using the A’s players’ past performance as a guide, he made reasoned arguments about how many runs they would actually score and allow. If they didn’t suffer an abnormally large number of injuries, he said, the team would score between 800 and 820 runs and give up between 650 and 670 runs. From that he predicted the team would win between 93 and 97 games and probably wind up in the play-offs. “There aren’t a lot of teams that win ninety-five games and don’t make it to the play-offs,” he said. “If we win ninety-five games and don’t make the play-offs, we’re fine with that.”
The A’s ended up scoring 800 runs and allowing 653. Not bad work by DePo.
I would assume that hockey teams or, at least the ones that are well run, go through similar processes. It’s harder to do this in hockey, for any number of reasons but, unsurprisingly, I do think that there’s a lot that can be gained.
A major distortion is the effect of the OT rules creating three point games in circumstances that aren’t particularly similar to NHL hockey: first four on four, then a shootout. This point should be of particular interest to the Oilers because they had some success in games that got to OT last year. The first question I’m interested in, as part of the broader question is this: how repeatable are results in games that get to OT?
I did some digging - the results are in the sidebar. What I’ve done is this. We have two seasons post-lockout for which we compare a team’s OT/SO stats with those they compiled in the following season. Obviously, we can’t yet compare the 2007-08 results to anything because we don’t have any 2008-09 results. The first chart on the left looks at points percentage. I’ve separated the 60 teams that I have to look at into six groups of ten, from best to worst, in each of two separate measures.
The first measure is games that went to OT/SO, as compared to the league average. What I did, for each of the 60 team seasons that took place in 2005-07, was divide a team’s OT/SO GP by the league average. For instance, in 2005-06, the average NHL team played 18.73 OT/SO games. The Oilers played 26. 26/18.73 = 1.39. That’s a big number - the 05-06 Oilers are second in the group of 60 in OTSO GP/LG AVG OTSO GP.
If you examine the chart, you’ll see that volume of OT/SO GP in one season appears to be a pretty poor predictor of volume of this games the following season. In year one, the teams in the group that played the most OT/SO games averaged about 24 each; the teams in the bottom pairing averaged abou13. In year + 1, neither of those groups was at either extreme; the spread between the groups that played the most and the fewest games respectively was about 3 GP.
Now look at the points percentage. The regression towards the mean seems ever more aggressive there to me. Assuming that all of those teams played league average games, the teams in the best group picked up about an extra 8 points over the teams in the worst group in the first year. In year + 1, the difference is down to about .6 points.
How does this apply to the 2007-08 Oilers? Well, they were outliers in both OT/SO GP/LG AVG and points percentage. In fact, of the 90 post-lockout team seasons, they rank third and second respectively. If both of those “strengths” of the 07-08 Oilers were in areas that seem to show a strong tendency to revert to average - as the “ability” to play OT/SO games and the “ability” to win them seemingly are - one wonders if, perhaps, last year’s Oilers weren’t quite as strong as the standings suggest.
Just for fun, I created something of a junk stat to try and measure the impact of the NHL’s point system on a team’s points. What it does is measure the difference between the points that a team actually got from OT/SO games and the points that they would have got, assuming (a) that they played a league average amount of OT/SO games, (b) had a .750 points percentage and (c) had a .500 record in any games that I’m now treating as regulation games.
To give an example of how this works, the 07-08 Oilers played 25 OT games in a league where 18.13 was average. They earned 44 points in those games. 18.13*0.75*2 (to convert from percentage to points) is 27.2 points. That leaves 6.867 games in which I’m treating them as having a .500 record for 6.86 points. Add it up and you get 34.067 points from games in which they actually got 44 points. That’s the biggest positive difference any team has posted in this junk stat in any post-lockout season.
The effect of wild variations in OT/SO results actually looks to be pretty small; applying my formula to every season since the lockout, 2/3 of teams are within +/- 3.5 points of 0, which represents playing a league average amount of OT/SO games with league average results.
If I’m Paul DePodesta, I’m looking at my team and thinking that it “really” got 78 points last year; if I want to realistically expect to contend for the playoffs, I think that I need to add about 8 wins to my roster. And that’s before I consider other things, like shooting percentages and injuries.
It’s going to be interesting to see what happens with the Oilers next year. If they put up 88 points again but do it with average OT/SO results, that would, to my mind, represent a great leap forward. Listening to Stauffer though, and reading some of the off-season comments from Oilers management, I get the impression that that might not be the common perception. Subject to my thoughts on shooting percentage and injuries, I think that overdelivery is going to be tougher than the Oilers might hope next year.
For those interested, the results that teams have piled up over the past three years are provided in the chart below. Flames fans must be frustrated.
Mudcrutch,
Nice Work. I have suggested on numerous occasions that the statistical probability of the Oil replicating this past season’s success in OT/SO is problematic for the team to a take a major step forward.
They will be hard-pressed to win 44 games; especially if management thinks they only have to add a couple players for an elite team
When did DePodesta’s parents split up?;)
Since negative results should be reported whenever possible, I’ll note that I was concerned about the throwing-together of the OT and SO settings so I tried to drill down into SO-specific data for year-to-year intercorrelations on the team level.
None were present, where I looked (converting figures in every case to standard deviations from the mean). Winning percentages, shot percentages, and save percentages in Year One for NHL teams do not appear to be helpful in predicting the same numbers for Year Two. The MOST impressive intercorrelation I saw was a paltry 8% r-squared for SvPct from ‘06-’07 to ‘07-’08. (I would instinctively expect save percentage to be the most reliable component of SO record between consecutive years, since the identity of the goalie will be the same in most cases and some noise from missed shots is filtered out of Sv%s.)
On a more interesting side note, the league save percentage in shootouts was .660 in ‘05-’06, .658 in ‘06-’07, and .657 in ‘07-’08, so neither the shooters nor the goalies seem to be gaining on each other as time goes by.
So in other words, while there is some minor evidence that Garon will continue to be better than average in the shootout next year, there is none whatsoever that the Oilers will play as many OT/SO games, or that so many of those games will be scoreless in OT, or that 10/83/89 will repeat their decent success rate in the SO.
Yeah, I’m not ready to install the Oil as NW Division favourites quite yet either.
As I noted yesterday, the Oilers winning percentage in 1-goal games was .641- good for 1st in the league.
Either we’ve got a team that can win one goal games, or a team that got lucky. I’m going to go out on a limb and say lucky.
Shootout wins would be considered one goal game wins would they not?
Great stuff. About what I expected regarding OT. And I wouldn’t expect even the Oilers management to count on OT points, though I think they’re probably pretty confident if the games go to OT. And I actually expect that to make at least a slight difference in how tight games are played. Which could give a slight bump in the number of OT games played, maybe a game or two.
The team will still be quite young. I will be skeptical about all predictions that do not factor gained experience into account, how many wins difference will the year make?
MC, your posts are beauties. Maybe you should post more on Saturday so I can get something done in a day.
Your 78 points could actually be closer 76 points if you factor into it that they were only getting .77 ppg in non-OT games (44 pts in 57 games).
Gotta love the NHL these days. Winning half the average number of OT games (18) amd 0.500 hockey gets you 27+64 = 91 points. You barely have to be better than average to make the playoffs. I don’t think its coincidence that the Oilers have a lot of OT games over the past 3 seasons. Its a good strategy for an average team.
Given that the Oilers have averaged more OT games (20.7) than other teams over the last 3 years and that Garon is an above average NHL shootout goaltender, I think its reasonable to say they will be above league average in again this year.
The 44 points in 57 games is where the Oilers should show the biggest improvement. The Oilers are better than a 0.386 club (44 pts out of 114 pts). The following factors (although cant be backed up mathematically), will IMO improve their non-OT record above 0.500
1) KLos constant win now attitude and willingness to add at the deadline
2) Improved PP numbers with Gagner, Hemsky setting up Sourays bomb (our PP was brutal first 25 games)
3) Penner actually showing up for training camp in shape this season and with chemistry with 10-83
4) Overall improved play from Gilbert, Grebeshkov (he was horrible first half last year), Gagner, Cogliano, Brodziak and Nilsson. Some of these players will decline but overall these players should improve
5) One of Morreau/Souray having a healthy season (just a hunch)
Jonathan: Does your stat “the Oilers winning percentage in 1-goal games was .641 - good for 1st in the league” include OT games? Because the Oilers won 19/41 games in OT and by my math only 6/41 by one goal outside OT. The 61% isn’t really a reflection of them getting lucky but rather that they played a lot of OT games.
Vancouver Canucks tried to bank on the same OT/SO “success” carrying over from 06/07 to 07/08. It cost Dave Nonis his job.
You can’t count on winning Wack-A-Mole every year to get you into the playoffs.
Jonathan: Does your stat “the Oilers winning percentage in 1-goal games was .641 - good for 1st in the league” include OT games? Because the Oilers won 19/41 games in OT and by my math only 6/41 by one goal outside OT. The 61% isn’t really a reflection of them getting lucky but rather that they played a lot of OT games.
Yes, that statistic includes overtime and shootout games (it’s from NHL.com, under goal-games). However, I’m going to disagree that it isn’t a reflection of the Oilers getting lucky- playing a lot of overtime games is still lucky, yes?
64% of all close games were won by the Oilers, and unlike a lot of commentators out there, I really believe that winning close games isn’t something a team can rely on- ie “clutch” scoring is largely (not entirely, but largely) a myth.
That suggests another research program: looking at whether records in one-goal games, including OT/SO games, hold up from year-to-year. If the Oilers play fewer OT games but win the close ones in regulation instead, that hardly leaves them worse off…
What a mishmash. You start out talking about RS and RA and then that’s the springboard for the OT record bugbear? I don’t get it.
Now if you could compare an Oiler’s team that had consistent results throughout all of last year, with only a few injuries, and with not a lot of young developing players, I think you’d have a reasonable chance of predicting next year from last, if…
you looked at GF and GA, like DePodesta. OT results, nonsensical at best, are not a predictor of success, nor LACK of it. Judging by your results we should be in OT less, so it should be less of a factor. Why are we considering it?
Our best comparison is from the last 30 games last year, yet we were still missing Horcoff. And Souray never played a single healthy game for the Oil.
Take those 30 games, add Souray, & Horc, place the kids on a year to year development schedule consistent with players of their calibre, and then let us know what results are predicted.
But let’s ignore the last 30 games.
If based on no change from last year, season numbers extrapolated to 82 games:
L1 - 75 goals
L2 - 42 goals
L3 - 50 goals
L4 - 30 goals (used Torres, not GX)
Defense - 45 goals, and that’s based on a Souray with a busted wing.
Comes to 242 goals with NO improvement by the kids, or Stoll, or Torres (both of whom under-performed their history).
Question is how many goals we will allow? I’m satisfied that with Garon installed as starter, a full year of Horc (so Stoll back to easier minutes), healthier Souray, defensive improvement from the kids, including a turnaround from Grebs, Gibby better used to an 82 game season, and better health fr Moreau, Pisani, & Torres, (3 of our best defensive forwards), we will be able to get our GA under 242.
We allow 20 less than we score, and we’re probably a 5-7 team in the conference, which would be 2 points out of third.
If so, this should be a playoff team. The critical factor will be the team’s confidence/poise, not the team talent.
If Souray plays 82 games, I personally think this team might score 280. And give up 300.
If Souray plays 82 games, I personally think this team might score 280. And give up 300.
That wouldnt be all bad; the brains over at HFBoards could talk about how Shawn Horcoff is finally a 90-point first line centre. Plus, Penner could hit 30!
A couple of things.
One - in marketing, when we see an aberrant stat, we consider two possibilities. One is that its simply an aberration, the other is that its a trend.
Second - your analysis makes the unwritten assumption, “all things being equal”. In this case, I’m not sure if all things are indeed equal. If they were, then I’d agree that eventually the Oilers would have to swing back to the median.
There’s a poster by the name of Bruce here abouts that made a pretty strong case last winter that the Oilers were working the system to their advantage. I tend to agree with this idea as I’ve heard from more than one source that the Oilers practice an inordinate amount of 4X4 and SO scenarios with Garon in net.
If you have a very young team that can compete, but doesn’t have the ability to close, then you would expect to see more OT situations. If that were your expectation, then it would make sense to get a great SO goalie (or maybe in this case you’d hope for one), and you’d practice the shit out of your 4X4 and SO. Thus, you’d be stacking the deck if and when it came to a SO. The SO is a skills contest. But like any skill, you can get better if you practice. And the 4X4 favors guys with wheels and lungs, which we’ve loaded up with.
I’m not saying MacT is a frickin’ genius. But he’s exceptionally good at working within the rules of the game. If the team is confident in OT, its because they have anticipated the situation. And by all accounts, all things are not going to be equal.
BTW - Thanks for the Oil-talk. There’s not many guys talking about the team right now. If you’re addicted to the ’sphere like I am, posts like yours are really appreciated. Keep up the great work!
What a mishmash. You start out talking about RS and RA and then that’s the springboard for the OT record bugbear? I don’t get it.
Well, it was kind of a lead in. Hockey’s tougher because straight goal differential doesn’t tell the truth anymore.
…you looked at GF and GA, like DePodesta. OT results, nonsensical at best, are not a predictor of success, nor LACK of it. Judging by your results we should be in OT less, so it should be less of a factor. Why are we considering it?
I agree with you. I don’t think that OT results are a predictor of anything at all. Hence the post. The problem though is that this Oilers team rode the OT results all year long. Without those OT results, they’re a lot further away from the playoffs than they looked. I kind of get the impression that management buys into that.
Our best comparison is from the last 30 games last year, yet we were still missing Horcoff. And Souray never played a single healthy game for the Oil.
Why is the best comparison the last 30 games? Just curious.
There’s a poster by the name of Bruce here abouts that made a pretty strong case last winter that the Oilers were working the system to their advantage. I tend to agree with this idea as I’ve heard from more than one source that the Oilers practice an inordinate amount of 4X4 and SO scenarios with Garon in net.
The Oilers got outshot 76-46 in OT and came out ahead 3-0. I think we just see things differently David - I see a team that got bombed at 4X4 and was probably lucky not to lose some OT games. I don’t believe that this was a result of strategy.
As for MacT just playing the system like a fiddle…I don’t see it. I’m a bigger MacT supporter than virtually anyone you’ll find but, and I find this odd, I’m about the least likely to believe that he has this power. If MacT is so good at getting the Oilers to OT, what the fuck happened in 06-07? The Oilers weren’t THAT bad prior to the Smyth trade but they weren’t any good at getting to OT. In 05-06, they were awesome at it. How do you guys explain this?
I’m more willing to believe that winning shootouts is a skill but Cosh’ numbers kind of suggest that there isn’t much to that, I think.
4 on 4 in regulation; Oilers get outshot 37-25; scored 2 and allowed 2. So this team, which you, David, think was a good 4 on 4 team, got outshot 113-71 when they were 4 x 4. I can’t square that with them being a good team; that’s horrible.
straight goal differential doesnt tell the whole story? Who had the best goal differential? Who won the cup? ya ya the wings are an anomaly. Allright.
I remember in ‘04 when Calgary was the absolute numero uno with one goal games. All thanks to Kipper i guess. They won a record amount of one goal games. Anyways, where did they go from there…
The critical factor will be the team’s confidence/poise, not the team talent.
I know it’s the first thing I look for in a successful sports franchise.
Thanks for the DePodesta link. I still look over his “The Genesis, Implementation, and Management of New Systems” piece from time to time. It’s a beauty.
As a few others have mentioned, the numbers are assuming “all other things being equal”. This is a convenient assumption for statistical reasons, but it doesn’t apply to the real world of the NHL. “All things being equal” assumes that the Edmonton Oilers have the exact same shootout abilities as the Toronto Maple Leafs, and that is obviously not true. (I’m not sure who is actually better between these teams, I’m only making a point that no two teams are exactly the same.)
Has it occurred to anybody that perhaps the Oilers are just good at shootouts? Shootouts are not a roll of the dice (a repeatable chance event), as they are being treated in this post, and therefore do not fit nicely into statistical methodology. The skill of the shooter and the goaltender obviously come into play. I know it sounds stupid to have to point this out, but the point nonetheless needs to be made, because it is largely being ignored here. You simply cannot state “all things being equal” Sidney Crosby and Zach Stortini each have an X% chance of accomplishing some task, as the skills of the players involved in the comparison carry a heavy weight.
As far as the Oilers repeating last years successes in OT and the shootout, I also think that they are unlikely to be as successful. However, I do think that they are likely going to be an above average team in this regard once more, so the fall back to earth won’t be as dramatic as many think. We still have a very good shootout/breakaway goaltender and we still have a good group of shooters, and these are the variables that determine the outcome of any shootout, not a random number generator.
The critical factor will be the team’s confidence/poise, not the team talent.
This reminds of the Seinfeld episode “The Chaparone”, with Kramer screaming “Poise counts” at Miss Rhode Island as she tunelessly grinds her way through “It’s a Most Unusual Day”.
“4 on 4 in regulation; Oilers get outshot 37-25; scored 2 and allowed 2. So this team, which you, David, think was a good 4 on 4 team, got outshot 113-71 when they were 4 x 4.”
That says to me a team which we acknowledge to be somewhat lacking holds its own on a 4X4 situation, despite being outshot (pointing to a better goaltender). That’s the idea. If you know you’re inferior at full strength, play to the advantage you can create. Hold ‘em at 4X4 and kill ‘em in the shootout.
Leading back to my original assumption, which is that in this case, all things are not equal. at which pure statistical analysis breaks down.
Not that I’m saying you’re wrong. Just that there might be another valid point of view that can’t be connected to a purely statistical breakdown. Or at least, maybe the numbers need some interpretation.
Strong work, Tyler.
If the Oil are hoping to replicate shoot-out record, and ride that to the playoffs, it’s evident that is a pipedream.
All this makes me think that the Oilers rebuilding plan is move slower than some fans think, which in my mind is an argument NOT to pick up short-term help for next year (please not even a hint of a Nylander/Souray type signing), because the idea should be to build for further down the road.
P.S. Souray bashers, yes his contract was risky, but his defensive play last season was strong. I’ve yet to see a shred of evidence from last season that it wasn’t, at least compared to other Oilers d-men.
If, in the unlikely event Souray is healthy for much of this season, we can expect strong defensive play from him.
P.P.S. Who is this “David S”?
Jonathan: I stand somewhat corrected on your stat. I was under the assumption that OT losses didn’t count as losses because the Oilers still got a point (they should be weighted differently however because losing and getting 1 point is obviously a lot better than losing and getting no points).
Playing a large number of OT games is a good strategy for a borderline playoff team - especially one that is good in the shootouts. Regardless, as MC pointed out the Oilers will need to a lot more points in regulation next year. Barring an early season collapse, I think they have the team to do it. IMO the Oilers going into next season are a lot better team than the one that was 14th in the west before the run.
Fine work, Tyler. But …
The Oilers got outshot 76-46 in OT and came out ahead 3-0. … 4 on 4 in regulation; Oilers get outshot 37-25; scored 2 and allowed 2.
The Oilers actually went 4-2 in overtime games, losing in Dallas and Minnesota, beating St. Louis (Penner), and CBJ-CHI-STL in succession on Cogliano’s astonishing “natural hat trick”.
In regulation 4-on-4s this season the Oilers outscored the opposition by the remarkable aggregate of 1-0.
I have not checked but have no reason to question your shots totals, which indicate the Oilers goalies had a dazzling Sv% of .982 in what is supposed to be a wide open manpower situation. That is not something that is likely to be replicated any time soon.
There’s a poster by the name of Bruce here abouts that made a pretty strong case last winter that the Oilers were working the system to their advantage. I tend to agree with this idea as I’ve heard from more than one source that the Oilers practice an inordinate amount of 4X4 and SO scenarios with Garon in net.
The system can be worked in two stages: 1) get to OT as frequently as possible; 2) win more than your share of the mini-games. Oilers have led the league in category #1 in two of the pass three seasons, and in #2 this past season. Getting to OT in the first place is rather easier to accomplish, since it works to the advantage of both teams to play a three-point game; in OT, when both are playing for that third point and only one will get it, one can expect a little more competition and a little less collusion.
The combination of playing the most OT games in the league AND having the best Win% in one-gola games is very remarkable. More OT games should mean more one-goal wins and losses both, and tend to moderate the percentages. The Oilers 2007-08 results are exceptional — surely 44 points from OT games are an NHL record under any points system and may stand for a good long while. I agree with Tyler that a bounceback is to be expected. Fortunately, I also agree with Sean that bonus points aside, the Oilers were a much stronger team at the end of 2007-08 than they were at the beginning, so (poor) regulation time results aren’t likely to be repeated either.
Playing a large number of OT games is a good strategy for a borderline playoff team
Is it a strategy at all? I can’t think of any ways that MacTavish could ensure his teams play a lot of overtime other than sitting back in the third period when the game is tied.
The Oilers actually went 4-2 in overtime games, losing in Dallas and Minnesota, beating St. Louis (Penner), and CBJ-CHI-STL in succession on Cogliano’s astonishing “natural hat trick”.
In regulation 4-on-4s this season the Oilers outscored the opposition by the remarkable aggregate of 1-0.
Given that the goal numbers work out and that I was gassed last night, I probably just made a mistake splitting them up.
Well the goal numbers work out cuz I “forced” one result (1-0 aggregate in regulation 4v4)from the other two, which are a matter of record (+5/-2 4v4, 4 wins 2 losses in OT decisions), but it can’t be any other way.
And I hasten to add, none of which invalidates your point(s) in the slightest.
Is it a strategy at all? I can’t think of any ways that MacTavish could ensure his teams play a lot of overtime other than sitting back in the third period when the game is tied.
Strategy may not be the best word but I don’t think its a coincidence that Oilers had the best SO record and the most SO appearances. A decision to play Stoll/Reasoner over Gagner/Cogliano late in a tie game could be considered “erroring on the side of OT” (obviously not for defensive zone faceoffs).
Getting to OT in the first place is rather easier to accomplish, since it works to the advantage of both teams to play a three-point game; in OT, when both are playing for that third point and only one will get it, one can expect a little more competition and a little less collusion.
How do you figure that it is in both teams benefit to add another point to the standings. Maybe if you are talking inter-conference games but in-conference that’s bunk.
If GD is important, Point Differential is more so.
Sketchy: it is to both teams advantage to go for the win when they already have a point in the bank rather than when they don’t. It is the fatal flaw to the Bettman Point scheme.
Point differential is indeed important, but it’s a matter of degree. In a three-point game, only one of your opponents will benefit, but any points that you gain yourself are gained against the whole conference.
For the sake of discussion, imagine an hypothetical stretch of games in the NW where every team plays each other twice, home and away. Let’s say the home team wins every game, but only the Oiler games go to overtime. The results would be: Edmonton 4-0-4, 12 points; all others 4-3-1, 9 points. The other teams would have all gained 3 points in their Edmonton games and just 2 points in games against each other opponent, but the Oilers would come out ahead due to gaining a disproportionate share of Bettman Points.
OF COURSE it’s better if you can win in regulation, and if you’re Detroit you do just that. But if you’re in the pack and play a lot of close games, it’s to your benefit if a disproportionate number of those games get decided in OT. There’s more points to be shared that way.
You barely have to be better than average to make the playoffs.
In a league where 16 out of 30 teams make it, isn’t that a mathematical necessity?
Well there ya go Bruce, OF COURSE its better if you win in regulation. That’s just what I said.
I do enjoy the rest of the analysis of the BPS. The only problem is you’re not guaranteed to win every game which goes to overtime or SO (or guaranteed that some other team doesn’t). In your example it benefits Edmonton if they are the only team going to OT - but they have no control over the games they are NOT involved in so I think that’s a great academic assessment which falls down in real life.
The only problem is you’re not guaranteed to win every game which goes to overtime or SO (or guaranteed that some other team doesn’t).
Right. And if you’re going to lose, it’s a hell of a lot better to do so in extra time.
they have no control over the games they are NOT involved in so I think that’s a great academic assessment which falls down in real life.
Well I said my example was hypothetical. Every OT game is beyond the control of the 28 teams not playing, and it’s a moronic system that rewards the two teams that can’t decide it in regulation. But it’s a logical conclusion that the more OT games a team plays the better off it will be, because the wins are treated like wins and the losses are rewarded as ties.
Oiler games were worth 189 points this year, those of the other teams in the Northwest between 180-184. How can that not be an advantage? It’s not so much an academic assessment as it is game theory.
I think the only argument that matters is that the Oilers have found a way to maximize their point take from an OT - SO scenario.
True, as they got their act together towards the latter part of the season they needed that tactic less. But I bet that they will still take a disproportionate amount of points this coming season from the total amount of games that DO go to OT. Maybe not to the same degree, but its apparent that a conscious decision has been made somewhere along the way to make sure those situations will not fall under the “all things being equal” category. That’s just smart coaching.
the more OT games a team plays the better off it will be, because the wins are treated like wins and the losses are rewarded as ties.
And the worse off you are the more OT games the opposition plays against other teams because their losses are treated as ties and yours are not.
I don’t really know a lot about the Oilers, but I’ll give an example from the early ’00s Red Wings which should help explain the issue with the BPS a bit more.
The Wings 3rd line was Draper-Maltby-McCarty. They could score a bit, but mostly they were great at keeping the other team from scoring. The Wings 4th line was Hull-Larionov-Robitaille. They could score a lot, but were a little old and not great defensively.
In the last 5 minutes of a game, if they Wings were losing, you’d want to see Hull out there. If no one scores, we lose, so we want to see a goal happen. If they score on us, no biggy, we lose anyway. If we score on them, we get to OT.
In the last 5 minutes, if the Wings were winning, you’d want to see Draper out there. If no one scores, we win, so we want no one to score. If they score on us we’re stuck going to OT, but if no one scores we win.
In the last 5 minutes, (and this is the key) if the game was tied, we’d want to see Draper out there. The reason for this is if no one scores we get to OT and then if Hull lets in a goal we still get a point. We can play it safe for 5 minutes, then play it lose for 5 minutes and the only difference is the losing team gets a point.
Yes, this does mean one conference opponent will also get a point. But since there are 14 conference opponents and only one of them gets the point we’re happy with this trade as long as we lose at least 1 OT game for every 14 we play.
This is true no matter who the teams are, which is why tied games should always go to OT in a ‘perfect’ world. Neither team should put their Hulls on the ice. It should be Draper against Madden for 5 minutes in every game, because both teams want to get to OT.
Of course, we’re talking about highly competitive players here, who would probably spear you in the groin if you suggested they intentionally not score when given the chance, but they’d still understand emphasising defense first at the end of a tie game.
I like your I like your premise, I like your logic, and I like the execution. Great work!
One might guess that when you DO take injuries and shooting percentages into the equation that Edmonton might expect to be slightly better off than 78 pts. My thinking without checking here is that the Oilers had the shooting percentages go their way moreso than not in 07/08 but that their rash of injuries compensates for this and then a bit more yet. Were I to make the guess right here and now, I would suggest 84-90 points would be a solid progression year for this young team whereas any less than that would be disappointing and any more would be serendipitous.
Yes, this does mean one conference opponent will also get a point. But since there are 14 conference opponents and only one of them gets the point we’re happy with this trade as long as we lose at least 1 OT game for every 14 we play.
And if all of your opponents play 2 for every 14 you still lose ground, to the tune of 1 point per 14 games or 5 to 6 points a season.
I understand the premise, even without the Detroit illustration, what I think many people are missing is that one team can only influence the outcome of one game, and two teams.
On paper it looks good, put in practice you’re playing some pretty bad odds because you can never influence any games but your own. It’s like the prisoner’s dilemma - I want to win and they want to win but we’ll both accept half a loss to ensure we get half a win, even though it can make us both worse off in the end.
In the name of seeing the forest through the trees a bit here… I think its an excellent point that the likelihood of the Oilers going into as many OT games as last season is somewhat unlikely. Given that, and the fact that they had such great success in a large number of OT games, then it is distinctly possible, if not probable, that one or both of those trends falter next season… leaving them to rely more on their abysmal regulation results.
This is a young team however, and one coming off of an injury marred season. The constants you can expect lies in the goaltending and coaching aspects, and I think Edmonton are ok in those departments. They tend to make the best out of turbulent situations.
Looking back to last summer, the Habs were coming off a poor season, they lost 2/3rds of their tough minutes line (Bonk, Johnson), they lost their PP leader (Souray). In the land of numbers, without looking, I’d assume most of the blogosphere pegged the habs to be a lottery pick team the past season… not so much.
Its really tough to get a pulse on the development curve of young players for next season, injuries and special teams results… all three will be significant contributors to the joys/sorrows of the Oilers in 08/09.
And no tyler, EDM being 30th on the PP for 1/4 of the season, and 3rd for more than half, doesn’t truely mean they are a 20th ranked powerplay.
I hope everybody’s keeping straight where exactly they believe in regression to the mean and where they don’t. I’d actually be willing bet that man-games lost to injury is a more persistent stat than any other mentioned in this thread, and the Oilers have now used this excuse twice in a row, but generally there seems to be a lot more faith in “We’ll be healthier next year!” than “We’ll get fewer OT/SO points next year!”
You’re off base Sketchy, and it’s precisely because you can’t influence games you’re not involved in. It doesn’t matter if every other game goes to OT or if no other game goes to OT. In terms of whether you want your game to go to OT or not, that is.
Given the choice between losing in OT or losing in regulation, the choice is clear. You want to lose in OT if you’re going to lose at all.
Given the choice between winning in OT or winning in regulation, the choice is again clear. You want to win in regulation.
However, you don’t know the outcome of the game until after it ends. So what you need to consider is the difference between winning in OT vs losing in regulation…
When you lose in regulation, you lose 1 point. When you win in OT, your opponent gains one point.
Now, there are 14 opponents to care about, right? The rest of the conference. In the first case, when you lose in regulation, you lose 1 point against every opponent. In the second case, when you win in OT, you lose 1 point against a single opponent.
Compared against your particular opponent it doesn’t matter which happens. The two are equivalent, you lose one point against them.
Compared against any other opponent however, it does matter what happens. You lose ground on them all when you lose in regulation. You don’t when you win in OT.
So again, with 5 minutes left in a tie game, should you play for a win or play to get to OT? Assume the outcome of the game is predetermined but we don’t know what it is. Without knowing the outcome of the game, would we rather it end now, or end in OT? There are 4 possibilities…
It ends now, and we win. We gain 2 points on the entire conference.
It ends now, and we lose. We lose 2 points on our opponent.
It ends in OT, and we win. We gain 2 points on the entire conference except our opponent, where we gain 1 point.
It ends in OT, and we lose. We gain 1 point on the entire conference except our opponent, where we lose 1 point.
Assume the odds of us eventually winning are X. Then, the end result of ending the game now is…
VS conference: (+2X points)
VS opponent: X(+2)+ (1-x)(-2) = +(4x-2 points)
If we end the game in OT…
VS conference: X(+2) + (1-x)(+1) = (1+x points)
VS opponent: X(+1) + (1-x)(-1) = (2x-1 points)
With some common winning percentages…
(50% to win)
Regulation
VS conference: +1 point
VS opponent: 0 point
OT
VS conference: +1.5 points
VS opponent: 0 point
(75% to win)
Regulation
VS conference: 1.5 points
VS opponent: 1 point
OT
VS conference: 1.75 points
VS opponent: .5 points
(25% to win)
Regulation
VS conference: +.5 points
VS opponent: -1 point
OT
VS conference: 1.25 points
VS opponent: -.5 points
What these numbers show is it’s strictly superior to go to OT when compared against the other 13 teams. Against your single opponent you’re better off having it end in regulation whenever you’re above 50% to win the game.
The key is that there’s only the one opponent, so unless it’s a situation where they’re 8th and you’re 9th that you’d even consider wanting to end it in regulation.
Remember, it doesn’t matter what happens in any other game you’re not involved in. You can’t control what happens in those games, they happen as they happen. All you can control is the outcome of your games, and where those free OT loss points go. All math aside, sometimes those OT loss points go to you, so you want to get there to get them. Only if you win an absurdly high amount of the time will you end up not getting your ‘fair’ share of those points.
I hope everybody’s keeping straight where exactly they believe in regression to the mean and where they don’t.
Just in case this applies to me as I see it applying to a few others I want to be clear that in the post I made I was operating under the assumption that OT/SO points, injuries, and shooting percentage would all regress towards the mean.
There might be an argument for injuries being less likely to regress than other factors, however, because reinjuring part of your body seems to be easier than getting injured in the first place. Rational yes, and significant, maybe. Who is to say.