THINGS THAT MAKE YOU GO UH-OH

The Edmonton Oilers are 7-1-0 after eight games and everything is coming up McDavid. Is everything as sunny as the won-loss record? It depends. When Connor McDavid is on the ice, things are gold. Off ice? Well…. We are too early in the game to be able to trust any of the numbers, but let’s have a look anyway to see how some of these things are trending.

MCDAVID’S IMPACT

One of the major stories this season is the team’s gap in quality when Connor McDavid is on and off the ice. Let’s use Corsi for and against at 5×5 to measure quality:

  • Oilers 5×5 Corsi for-against: 323-361 (47.22)
  • McDavid Oilers 5×5 Corsi for-against: 128-116 (52.46)
  • Without McDavid Oilers 5×5 Corsi for-against: 195-245 (44.3)
  • Source

That is a massive gap, and frankly leaves the Oilers badly exposed in case of injury. One of the things Peter Chiarelli was counting on after trading Taylor Hall was a second scoring line developing. That hasn’t happened yet, and I do believe we should monitor Leon’s handling at the first sign of a scoring slump by the Oilers. I suspect he will get moved up to the Nuge line in a quick damn hurry.

SEASON OVER SEASON NUMBERS

I am going to write these categories with the 2016-17 numbers and then put (2015-16 numbers in brackets).

  • 5×5 Goals for Percentage: 60.0 (44.3)

Well, that is a spike for sure. Edmonton has gone from the bottom (No. 29 a year ago) to near the top in this category. The question we have to ask ourselves is in regard to sustainability, and there is some evidence that luck (Tyler Pitlick 3 goals, Nuge 0 goals) has played a part in both directions.

  • 5×5 Shots-For Percentage: 49.1 (48.5)

This is a slight improvement, don’t think we can trust this as being an actual reflection of improvement—we need to wait until 20 games are completed to trust these things. It looks good, trending well, but beyond that, not much to see here.

  • 5×5 Fenwick For Percentage: 47.8 (48.3)

A slight downturn, certainly not anything to fret about at this time.

  • 5×5 Corsi For Percentage: 47.3 (48.8)

This is the metric I follow closely as the season rolls along, there is a tie between successful teams and good numbers in this category. Edmonton has been ahead in several games this season, so score effects are a factor—but you would like to see this number edging closer to 50.0 percent as the season rolls.

  • 5×5 Shooting Percentage: 9.73 (7.17)

One of the things we are seeing this year is Edmonton doing well in areas where regression is likely. Shooting percentage is one of them, as hovering around 10 percent as a team isn’t something Edmonton does routinely. Shot volume is a favorite phrase of Todd McLellan, I believe the Oilers would have to increase the volume in order to score at current rates throughout the season. When we look at this after game 20, this number may be closer to last year.

  • 5×5 Save Percentage: 93.75 (91.53)

Cam Talbot has been terrific this year (did you see that stop last night late? Lordy) and that performance is reflected in these numbers. Edmonton has had an easier schedule in October, with plenty of home games, short travel and adequate rest between contests. There is a good chance this number slides in November due to all kinds of factors, including luck.

  • 5×5 PDO: 103.5 (98.7)
  • Source

PDO is a fascinating stat discovered by a brilliant Edmontonian while bird watching in the Yukon. His name is Brian King, and his little statistic is often overlooked for one reason or another. Gabriel Desjardins framed PDO and its effectiveness perfectly in an article from 2011, and I am going to borrow liberally from it here. Gabe’s main points:

  • Shooting percentage is primarily luck-driven
  • Save percentage is primarily luck-driven
  • Source

All PDO does (the metric, not the guy. The guy drives around a lot) is add shooting and save percentage together, and tells you how badly the crash will be when luck comes to town (or leaves it). If you are below 100 percent PDO, chances are a market correction will improve fortunes. Well above it? Look out below.

We wait.



  • KenBone18

    As much as I hate to say it here are the stats that matter:

    – Oilers 7-1; have 14 points

    That’s all that matters. You can dissect the minutia and talk about this or that, but at the end of the day, all that matters is win/ loss.

    It’s like going to your doctor, and your doctor says – “Well Kevin, I’ve made your blood work looks good, but you’re still going to die an early death”. Who cares what the numbers are, if you’re not going to live any longer?

    Who cares if you win (lose) the corsi, if the points aren’t there? Who cares if you win (lose) the Fenwick? Who cares if PDO reverts from 103 to 100?

    All hockey fans, coaches and GMs should care about are wins / losses.

  • godot10

    PDO is a valid zeroth order advanced stat.

    It should eventually be superseded by a proper calculation of expected goals for and expected goals against, based on shot location and shot type, and the deviation from expected goals for and expected goals against should result in a slight improvement on PDO.

    PDO assumes all shots are of equal quality. A proper calculation of expected goals for and expected goals against would give a better indication about how much a a run of good or bad fortune a team is on.

  • I LOVE the early season wins! That said my biggest concern right now is the team seems to be working at almost optimum, which is rarely sustainable in contact sports. If McDavid get the flu or Talbot stubbs his toe in the showers we’re in a heap of trouble.

    Vancouver played a neutral zone trap and almost had us, probably would have if not for McDavid. You can be sure other teams were taking notes. A guy earlier in the thread mentioned it – we’re generating most of our chances off the rush. When teams adjust to choke the rush off we’ll have to rely on a decent PP. Problem is McDavid can’t play all the minutes on all the PP’s.

    Don’t get me wrong. Overall we’re better. But I can’t help the feeling that in a modern game where analytics and video are so prevalent it’s only a matter of time before teams figure it out. I honestly think McDavid is masking the true gauge of this team.

    • ¯\_(ツ)_/¯

      I honestly think McDavid is masking the true gauge of this team.

      He’s the team’s captain and leader.

      How can he be masking the true anything of the team? I think the word you’re looking for is “exemplifying,” not “masking.” Similarly, Talbot wears a mask, but isn’t masking the D, he exemplifies the D.

      • No. What I’m saying is that McDavid is so good and Talbot is playing so well it’s masking the true sense of how good this team actually is. Only when Talbot cools off (he will) and teams figure out how to defend against McDavid (they will) or he has a stretch of “normal” (inevitable) will we see what we’ve actually got.

        • Dr. Merkwurdigliebe

          McDavid playing well and Talbot playing so good means we have a great number one centre and a good goalie. Take any team’s top centre and goalie away from them, and there will always be a drop off. And for the record, I think McDavey will only get better.

          • Admiral Ackbar

            The argument is that in that scenario, the Oilers would be worse off than other teams that’d lose their 1C and 1G. It’s a relative comparison. Is there ambiguity in the concern? Yes.

          • Dr. Merkwurdigliebe

            Maybe, but I like the fact we have Nuge and Draisaitl as 2 and 3. Those are two pretty high-end draft picks if you remember. Too much hand-wringing from a fan base that’s a little jittery (for good reason). For now, enjoy the wins! Season is usually over by now, but the Oil are first in the west! That’s awesome.

        • ¯\_(ツ)_/¯

          when . . . teams figure out how to defend against McDavid (they will) or he has a stretch of “normal” (inevitable) will we see what we’ve actually got.

          Got it. And when that happens . . . in 2024, I hope the masking you speak of has resulted in 4-5 Cups. Until then,it’s probably a safe bet that McDavid’s for real.

        • GK1980

          Teams will never be able to figure out mcdavid . His speed is in another level and his hockey iq is off the charts. His skating technique is beautiful and will change the way this game is played and how players skate.

          The past 20 years hockey all over the world has been focusing on defending. The offensive players are adapting to this suffocating method of playing hockey and a new generation of players has emerged. Speed kills D and we are starting to see this. Look at all the young kids emerging. Larkin, Mathews, Eickel, McDavid. They all play with speed and have outstanding puck management abilities.

          The trap is dead and is no longer an effective strategy to win, ask New Jersey and Minnie.

          • That “dead” trap almost beat us last night. Do you not realize how close that game was? Whether you want to believe it or not the Oilers basically have a rush game and a neutral zone trap is the best way to counter it. We’re going to see alot more of it in the future.

          • Reg Dunlop

            Are you referring to the trap that limited McDavid to only 2 clear cut breakaways? And if the Oil have basically a rush game how do you explain the Oil playing keep away in Vancouver’s end after the canucks pulled the Gump?

            I think many here are lost if there is nothing to gripe about. Hopefully Larsson’s game sours so everyone can resume lamenting Hallsy-baby’s departure.

          • I try to look at the whole game, not the 62 seconds in whatever period the Oilers FINALLY established possession.

            Hey look. All I’m saying is Connor isn’t a god and maybe the other lines aren’t nearly as good as they need to be over 82 games.

            I know it’s hard but this isn’t the Gretzky Oilers (they had stars down to the third line) and you guys are already painting them as cup champs.

          • GK1980

            Yup, the trap is working well for Vancouver. Point made. You are correct in that the oilers score more goals off of the rush. I would like to see more cycling down low, of course we have all been wanting to see this for a decade.

  • The Goalie 1976

    Hi Lowetide,

    Thanks for the article, I think the ON readers are giving you a hard time for being negative. I encourage you all to check out the below article from Lowetide’s blog. He does a brief write up on each line’s performance, from his ‘by-eye’ method. Also posts the Corsi, and the newish stat HD SCA (high danger scoring chances against). It’s really the combined global overall method us non-stats guys would prefer.

    http://lowetide.ca/2016/10/29/rainbow-rising/

    Myself, I am not a corsi guy, as being a former goalie, I can attest that not all shots are equal in their difficulty, and I’d rather have to save 5 good perimeter shots, that 1 shot from the slot.

    Lowetide is always pointing out the ‘kill floor’ which is high danger chances in the slot, and this appears to be a coaching style that is deriving success.

    Anyway, my point is this particular Lowetide article may be corsi heavy, but I think everyone can see the game plan is to focus on taking high danger shots on the kill floor, and defend ferociously the kill floor in front of Talbot.

    I believe this to be a sound game plan, no matter if the corsi says the Oilers lost the possession battle, which they did yesterday.

    If you check out the link above for Lowetide’s article, he points out the Oilers won the scoring chance stat, and he also posted this shot distance stat:

    Average shot distance:

    VAN 41 ft
    EDM 29 ft

    No, that’s not a typo.

    I believe what the Oilers are doing is a effective and sustainable strategy for this team, as you need good goaltending, size, speed, high skill, and defenders that can defend (Russell) bend and not break, then skate and make an outlet pass to the skill forwards.

    It would appear PC went out and built a team that McLellan can win with. (we are not used to this fr the old boys club times)

    Also to note in Lowetide’s defense, check out his ‘by eye’ notes on Nurse and Russell. Although he is a corsi guy, he can still recognize the progress Nurse is making now that the team depth is able to put him in a position to succeed (3rd pair) and Russell’s ability to defend like a bugger, and bend and not break.

    I will always be mildly interested in corsi, mostly just for curiosity sake than anything. But Lowetide’s articles also note the high danger chances for/against, that really make players like Russell, Sekera, Nurse and Larrson shine with their ability to push shots to the outside and limit the opposition to low danger chances.

    All good stuff

    GO OILERS

  • C’mon, people. Saying there are things that can improve isn’t being negative towards the team. LT is pointing out some numbers to work on, but nowhere did I see this as a negative article.

    The last 10 years have some of you guys a little too much on edge, not that I blame you.

    • Shameless Plugger

      I get what you’re saying, however remember a couple years ago when those inferior hockey players from southern Alberta were routinely getting smashed on the analytical numbers yet managed to make the second round of the playoffs?

      Numbers are great and all but don’t loose sight that this team is wayyyyy better than any we’ve seen in the last decade.

      I suppose it’s the incessant search for anything negative to write about that’s bothersome. Especially when things are so positive right now.

      Posters on here get labeled constantly as trolls by the bloggers/writers of this site. In my eyes this article is the exact thing we get called out for. Can’t we have at least a month of enjoying it before we disssect the most trivial of things ie. Corsi/fenwick/zonestarts/ quality of competition etc. etc……

      • I get the Flames thing for sure. That said, guys like Kent Wilson were talking about how those numbers weren’t sustainable over the long term and the next year Calgary crashed and burned. I think it’s just raising awareness more so than looking for something negative to write about. That’s not LT’s game.

        Like I said, I saw this as areas where the Oilers will need to improve to keep this beautiful run going as opposed to looking for negatives. If you’re looking for some positivity, feel free to re-read my wrap up. I was absolutely beaming last night.

        • Shameless Plugger

          I read the wrap up and everything about the oilers I could find last night. I can’t get enough oiler right now.

          I don’t think I’m oblivious to the fact there can always be improvement in anything all the time. 100% is next to impossible to attain no matter the topic of discussion.

          I suppose maybe I don’t fully understand analytics and that’s on me to get better at that, it’s just that the Oil are 7 and freaking 1 and looking good doing it.

          Nitpicking after ten years of futility just rubs me the wrong way right now.

    • Admiral Ackbar

      The arrogance of some of the posters on this site is getting alarming. The mere mention of something not in line with their beliefs, they attack the writer instead of breaking down their argument.

      I think the validity of 8 games worth of data is questionable but there is plenty of evidence that the middle 6 aren’t doing what we hoped they could do. It’s a perfectly reasonable article with some statistical evidence. There is absolutely nothing wrong with that.

      Some of these infantile posters need to be moderated. Criticize with a counter point, not uncivilized attacks.

  • Muddy

    I know McJesus is hands down our “best” player but how fricken good is RNH? Seriously turning into one of the most rounded players in the NHL. I notice McDavid with the puck but I always seem to notice Nuge without it. So glad he wasn’t the tradable player last summer.

  • stonedtodeath61

    Update from Ottawa. Anderson, who was away from the team on personal matters will join the club in Edmonton AND MAY START. Turns out Andy and his family has received news that Andy’s wife has cancer. Prayers and thoughts with him and his family.

  • Petrolero

    Honestly Lowe tide stop using numbers you don’t understand. Sometimes I think Dallas Eakins is writing these articles.

    Pdo is only useful over large samples like an entire season, more than 20 games and definitely more than 8.

    Score effects do play a factor, a big factor. if you’re stubborn about using putrid corsica and Fenwick, and nothing else, at lease use the adjusted Stat for score effects.

    I’m not a fan of David staples at the journal but at least the guy has the intelligence to record chanced according by quality. Hitchcock said it in his post game that the total shots was misleading because they were kept to the outside.

    • Lowetide has been using advanced stats longer than some fans have been watching the games. I am pretty sure he has some understanding of the numbers. You don’t have to like the implications, but they imply what they imply.

        • finn_fann

          I do want to say that this is a great read, which I haven’t seen before. That said, we’re talking team PDO in this case, and not individual PDO. Some teams have shown the ability to maintain an above-average PDO (mainly when they have an elite goalie), but banking on that being the case with the Oilers when they have no proven track record of being a plus team in this stat is a pretty high-risk bet.

      • Gadgets

        Thats the problem, he uses stats but has very little actual understanding of their proper use. Case in point, he constantly reviews single game corsi. No one with a real knowledge of analytics would bother with corsi over a single game. Its about as useful as looking at a single shift. He also admits that he pay more attention to corsi when every analytics guy has clearly stated that there are better metrics to use.

        • finn_fann

          Corsi has been fairly well-validated on a team level for predicting where a team ends up in the standings at the end of the year. There are many other stats out there, but things like HDSC are relatively new and don’t have the historical data (yet) to know what their predictive value is. I see nothing wrong on Lowetide’s part with being hesitant to jump on the shiniest new stats.

          Also, I don’t see Lowetide making any definitive claims based on the numbers here. All he’s doing is pointing out that the PDO is high and the corsi isn’t quite marching in lock-step with where we are in the standings. He says a lot of things along the lines of “you would like to see these numbers be better” (which is a fair point, imo), but he never says that the oilers are crap, as many posters are making it out to be.

          … just out of curiosity, what stats exactly are recommended by “every analytics guy” over corsi?

    • finn_fann

      The fact that you say wait 20 games to look at PDO shows that you do not understand this stat. The whole point of PDO is that, in the long term (say after 20 games), PDO tends to regress to 100. In the short term (say 8 games where you go 7-1-0), if you’re rocking a PDO of 104, then you are playing in good luck that is inflating your results.

      In the short term, a team can score on 12% of their shots, but a lot of work has shown that the same team is just as likely to go on a stretch where they’re scoring on 6% of their shots. You could have the same team and the same number of shots, but the number of goals scored would vary 2-fold.

      I remember the Wild making it to the playoffs on a PDO-fueled streak, as well as Colorado and Calgary. IIRC, none of the made the playoffs the year after. There always seems to be one team that lucks its way into the playoffs, and I am all for it if Edmonton can be that team this year, but I’d much rather they show improved corsi and fenwick that actually predict sustainable success.

      As far as corsi and fenwick are concerned though, I think it’s too early to worry about what these numbers show at this point. In contrast to PDO, you DO need to wait 20 games at least (if not 30 or 40) to make predictions with any sort of accuracy. There is also the fact that we’ve been playing with a lead, and score effects do matter. You can look at 5×5 tied corsi, but the sample size at this point would be even smaller and quite unlikely to give you any meaningful information.

      Tl;dr Edmonton is playing in good luck and is not going to go 7-1 the rest of the way. If the stats continue to look this way at the 20-30 game mark, I would be concerned. But at this point the Oilers sure look like an improved team to me.

        • finn_fann

          You mentioned in your first post that PDO needs to be looked at after 20 games, if not an entire season. I was pointing out that PDO is often used to point out whether the results over a stretch of games was driven by a big swing in shooting/save percentages (aka good or bad luck). Over the course of a year, this tends to regress towards 1.000, although you do see some teams ride a favorable PDO all the way into the playoffs.

          I think we both agree that a larger sample size is needed before corsi or fenwick are of any use.

          • Petrolero

            Just make sure you ead properly fore you reply or you look dumb. I wrote: “Pdo is only useful over large samples like an entire season, more than 20 games and definitely more than 8.”

            You basically bi t ched about something I didn’t write and then said the same thing I did. MORON.

          • finn_fann

            I honestly don’t know how to respond to this. Let me put our quotes side by side.

            I said: “You mentioned in your first post that PDO needs to be looked at after 20 games, if not an entire season”

            Your quote: “Pdo is only useful over large samples like an entire season, more than 20 games and definitely more than 8”

            Sorry if I didn’t copy and paste your words, but these mean essentially the same thing. In the mean time, do you have anything of actual substance to add to the discussion? Why does PDO have to be measured over a season? Is there a reason that you feel it shouldn’t be used to look at an 8-game stretch?

    • The game consists of tracking how many goals are scored, then awards points based on who scores more goals. Hockey is a numbers game at its core. Calling something that people get advanced degrees in “for losers” says more about the the insecurity of the person saying it than anything else.

  • BlueHairedApe

    I went to hockey analysis and had a look at last seasons numbers and it’s impossible to find any correlation with a teams PDO numbers and where they ranked in the standings at the end of the season. For instance Pittsburgh was ranked right alongside Arizona and Colorado and we all know there’s just a bit of a difference in the way those teams played last year. The ‘magical 100’ as it were is nothing but a unicorn and has no bearing on how these teams played or how it dictated who the eventual Stanley Cup champions were. If Las Vegas relied on these numbers I’d be willing to remortgage my house.

  • Masma

    I know there’s a
    -scoring award
    -defensive award
    -most valuable player award
    -most regular team points award
    -winning the cup award
    -etc.

    didn’t realize there’s a corsi award . personally corsi means crap when the team is winning and guys are playing hard every shift .
    I’m not planning a parade but I am planning to watch the next game something that didn’t always happen

  • Deke Rivers

    What are the numbers for lines 2, 3, & 4 for both years. Is this year an improvement across the board with mcdavid removed from both years? We knew what we were getting this year from mcdavid. The question is did the other lines improve?

  • Admiral Ackbar

    Great article, Lowetide, though I’d like to see the same analysis applied to all NHL teams. A lot of teams will play a lot worse when their #1C isn’t on the ice, no?

    These numbers reflect an extremely small sample size (8games) and a lot of the Oilers Corsi/Fenwick can be written off to scoring effects.

  • Harry2

    Man the stat geeks just never give it a rest. When the stats show what they want them to they gloat. When the stats show what they dont all of a sudden its not sustainable.

  • camdog

    Good article from ANDREW BERKSHIRE at sportsnet,

    “What may be even more important is that the Oilers are currently fifth in the NHL in high-danger scoring chances per 20 minutes played at even strength, up there with the elites, and they’re allowing fewer high-danger chances against than average, a huge change from the typically porous defensive teams we’ve seen in Edmonton.”

    http://www.sportsnet.ca/hockey/nhl/driving-oilers-hot-start-sustainable/

  • Hadley

    I can’t wait for the “luck” stat to start showing up on the back of hockey cards. Will it end up as a fantasy hockey category too?
    Luck is very evident in these advanced stats if the numbers don’t work well with wins and losses.. “team A lost in the Corsi to team B. But team A still won..Must have been luck, no other reason why that could have happened”
    Well you have to be good to be lucky.. and the more you practice the luckier you get.
    Can someone please tell me what pleasure they get in studying these stats? Is it to one day rub it in someone else’s face that you were right that the team would lose? I’m asking this in a sincere way. I have just never understood it. I get my rocks off for W’s. Are you going to brag when your team loses but wins in the advanced stats category? I sure hope not. Enjoy the wins. I understand blogs need to be written and money needs to be made, but this is just sad and boring to read all the time.

  • Speaking of luck and good/bad bounces…had Baertsche scored on that open net it would’ve been a different game.

    So I see it happening, I see the effect of “luck” in these wins.

    But I’m still happy, because they don’t look like a tire fire in their own zone. The D looks calm and relaxed with the puck as they, with deliberation, move the puck up ice.

    I also see progress in their cycle game, which just needs a few more weeks of chemistry building, where these cycles will result in shots on net.

    Progress!

  • whateverhappenedtoearledwards

    I have often wondered how the Oil I\of the glory days would have looked in the advance stats world. The glory Oil were amazing at cashing in on scoring chances with the high quality shooters they had. They also didn’t waste a lot of shots, their shot totals were often not as high as one would have expected for the number of goals scored.

  • Joy S. Lee

    Luck is more evident in today’s game than ever. The reason for this is that time and space have been significantly reduced due to vastly improved skating ability and coaching systems. Also, today’s sheets of ice are torn down and rebuilt almost daily.

    As a result, puck battles are condensed, bodies and sticks flying, and the puck bounces unpredictably. Today’s younger players are accustomed to this, and play the game accordingly – they anticipate the chaos of the bouncing puck better than previous generations.

    My point is, the luck of a single bounce can affect the outcome of a game, particularly if it’s a low-scoring affair. Good luck measuring and quantifying that. Let’s not go down that dull road. I think I’d rather just watch the game.