Fall in Edmonton is a glorious time to be outside. Leaves are falling in colours red and gold, everyone and their dog are out and walking and enjoying these moments. This morning, it’s just about freezing at my house, wouldn’t be surprised to see some frost on whatever you left uncovered last night.
We’ve just dipped our toes into the 2024-25 season, and the Oilers joined the season midway through the Flyers game. Under Kris Knoblauch, the team often spends this period tearing up the blacktop, winning many games and posting points as far as the eye can see. The first road trip is here. It’s a significant test.
The Athletic article looking at Ty Emberson is here.
WHAT TO EXPECT FROM OCTOBER
- At home to: WPG, CHI, CAL, PHI (Expected 3-1-0) (Actual 1-3-0)
- On the road to: NAS, DAL (Expected 1-1-0) (Actual 0-0-0)
- At home to: CAR, PIT (Expected 1-1-0) (Actual 0-0-0)
- On the road to: DET, CBJ, NAS (Expected 2-1-0) (Actual 0-0-0)
- Overall expected result: 7-4-0, 14 points in 11 games
- Actual October results: 1-3-0, two points in four games
- Oilers in 2024-25: 1-3-0, two points in four games
I have the Oilers hovering around a standings point-per-game over the rest of the month, that would mean a final October record (including what we’ve seen) of 5-6-0, 10 points in 11 games. Nothing tragic, and we wouldn’t be able to say the gales of November came early. However, this team going 4-3-0 over the next seven isn’t guaranteed.
I like to look at Corsi and Fenwick and individually the rel numbers. I don’t think anyone does that anymore, but for me until Puck IQ gets rolling it gives us a strong view of reality. We have to take score effects into account, and that’s for sure with this Oilers team after four games. Having said that, the team’s five-on-five Corsi (62 percent) and Fenwick (58.6 percent) are tops in the NHL at this time. Expected goals (57 percent) at five-on-five is No. 3, and the actual goal share five-on-five (5-10) is 33 percent and ranks No. 29. We can see the first four games well enough in these numbers, but I would suggest that the math is misleading due to score effects. The Oilers are very good, but also flawed.
I’m not thrilled with the results from the Darnell Nurse pairings so far, but am pleased with the process. Kris Knoblauch’s choices for No. 2 RD can all move the puck, jump into the play and get back to where they once belonged for puck retrieval. I’m tempted to fade the results in the first three games for Nurse and partner because he looked for more able in Game 4. He played two preseason games, folks, it should be no surprise that an extended training camp took place during the regular season.
Corsi Rel five-on-five shows Evan Bouchard shining like a diamond, he’s shooting the puck a lot early on. I think he’s started a little slowly, too, meaning the best is yet to come. Brett Kulak and Mattias Ekholm also perform well Rel. The downbeat fellows are Ty Emberson (just underwater), Troy Stecher (a little deeper but you could save him by reaching out your hand), Travis Dermott (you’d have to walk into the ocean a few steps and physically immerse yourself into the water in order to save him) and then Nurse (you hope he’s got an oxygen tank).
I don’t think this team is going to be tight defensively ever. If you are looking for the Jacques Lemaire Devils, or the tightness of Earth, Wind & Fire, another NHL team may be to your liking. The Oilers, at their best, are like Crazy Horse. Loose and raw, at times at the outer edge of what might be called music, but out of this world majestic in moving you emotionally via song. That’s the Oilers. High, high high ceiling, low, low, low floor. Tonight, it’s Nashville.
The Thursday edition of the Lowdown hits at noon, Sports 1440. Bagged Milk from Oilers Nation will help us preview the game tonight, and we’ll have the latest from Nashville re lines and pairings. It’s a short but important road trip, I’ll be looking for Troy Stecher alongside Nurse and then it’s three chord rock and wait for the mind blowing solos. I’m at Lowetide on twitter, in the comments section here and on the Sports 1440 text line at 1.833.401.1440 directly. We can be heard at sports1440.ca; iHeartRadio; Radioplayer Canada, we tweet out the show after it’s done and you can catch us on Apple and Spotify.
The main problem I have with stats like Corsi and Fenwick is I forget their definitions.
I know they mean something!
Haha, that’s me too. I don’t speak fancy stats
They tell us where the puck is when groups of players are on the ice. That’s about it. They seem fairly accurate in this regard since SN (etc) have been putting actual OZone time stats up on the broadcast. There has been a pretty strong correlation–although we don’t get the final OZ time stats, Oilers always seem to be leading thus far this season when they are posted, confirming the two shot attempt stats.
The problem comes in determining how much effect any individual player has in determining where the puck is. A lot of license can be deployed in this regard.
I read you post variations of this, often, and wonder if it’s more of a meme or a poke at the Twitter crowd?
Because folks around here use Corgis, DFF, and -rels often.
I don’t think there have been any score effects. In the Winnipeg game the shots were 9-4 Oilers when Winnipeg scored their first goal and 9-8 after the second goal. At no point in that game were the Jets outshooting or outchancing the Oilers. The Chicago game was 2-1 going into the third, so at most when the game got to 3-1 the oilers maybe had 5 shots all season that can be attributed to score effects…the Flames game was tied 1-1 going into the third and the Oilers were outshot when they were down..
For me, Bouch is having his frustratingly typical early season puck issues. Lots and lots of puck bobbles and whiffs that are leading to high danger chances against. This is at least 3 years in row and is now a real thing. What is heartening is the knowledge that he will indeed get those bobbles out of his game (for the most part) and will soon be that elite 2-way d-man he’s proven to be.
He’s shooting alot but he also has something like 28-29 shots blocked – in four games – that’s an astonishing stat.
What source are you using for his blocked shot totals? Hockey reference only has him at 5 over 4 games: Evan Bouchard Stats, Height, Weight, Position, Salary, Title | Hockey-Reference.com
Believe the BA stat is blocked attempts which is 28. He’s not been shooting with purpose, his famous headlong passes are also giving similar vibes. His calibration is off.
Disregard, I’m misreading as him blocking shots vs his SOG’s getting blocked.
I don’t think Stecher is a longer term answer at 2RD, and I’m not positive his body can hold up through 82 games (let alone playoffs) with his style of play, but early returns on Nurse/Stecher are heartening.
This duo has now played 58 minutes together (last year and this) – most metrics are 50% across the board, except the high danger Corsi is 11-6 leading to a 60% expected goals and….. 4-0 real goals.
50 of those minutes, and all four goals scored, are without McDavid!
From my recollection, I’m not sure the Oilers’ numbers benefit very much from score effects.
Maybe in the Winnipeg game but I’m not sure the Oilers pressed in the 3rd period of that one and benefited from a sitting back Jets team (I honestly can’t remember).
In the Hawks game, I recall the Oilers chasing the score but much of that was tending and they were leading the possession game throughout.
The flames game, the Oilers were ahead for much and the flames pressed in the 2nd half of the game.
In the Flyers game, the Oilers found their game in the 2nd, yes they were down, but I don’t think this was a case of either team changing their game due to the score but the Oilers finally “showing up”.
Maybe I’m misremembering how the flow of these games went.
Against Winnipeg the Oilers lead the shots, hits, and faceoffs from start to finish. Was a classic mostly goalie defeat.
Per Stauff, same lines and pairings as last game and Pickard gets the start!
Nice to see them getting Pickard engaged this season. They seemed averse to giving him starts last year. I think him getting minutes early in the season will pay off down the road, with him being well primed (and the team being more comfortable/familiar) and Disco Stu being more well rested for the playoffs.
Bouchard has 7 sog, his partner Ekholm has 12. The eternal punching bag Nurse has 9, Kulak the defender people always forget about when they say Edmonton doesn’t have 4 defenders has 7.
Bouchard needs to do a hell of alot better job doing what he’s always been known for, getting his shots through to the net.
Hyman with 5 sog, with an absolutely selfish penalty last game. Of all the players that need a message sent to by the coach, Hyman needs to be pushed down the lineup until his play begins to warrant the best RW job in the entire league.
Yes, Hyman scored 8 billion goals last year. Yet his team 1-3 this year, it’s a what have you done lately league, and what Hyman has done lately is contribute to losing.
Hyman is the last player at the moment I want to see running alongside the glimmer twins.
Draisaitl with a big game tonight.
Hyman didn’t increase his lines production, he just took goals from Connor (not trying to score last season) and Nuge, who returned to normal levels. The line last season scored 3 more goals than the previous. To me he needs to forecheck, hit, get after anyone bothering Connor even if it’s not fighting, and be back fast with his speed and be reliable defensively, which he isn’t always, to be on the first line
Connor has had better help than Leon, but to me not ideal either. Young Perry or healthy Kane and a strong two way is what I think. Nuge comes and goes as a two way player at this point
I don’t think Connor cares who scores, just that goals go in. And between 22/23 and 23/24, Connor’s 5 on 5 on ice GF/60 and points per 60 both jumped substantially. His 5 on 5 GF/60 went from 3.44 to 4.38 (with no change in GA/60) – highest in his career; P/60 went from 2.71 to 3.5, the second highest of his career and barely behind his best season in the 56 game Canada only division. All data from Naturalstattrick.
Naturalstattrick also shows that in each of those seasons, McDavid’s production and goal share was way better when he was in the ice with Hyman than on the ice without Hyman. Finally, In 22/23 RNH played way more without McDavid than with him and didn’t become the full time LW till last season.
McDavid actually attempted as many shots 5 on 5 in both seasons but more got blocked and more missed the net. Defenses probably played him differently because of the 60+ goals the previous year, but according to NHL edge his shot velocity was down almost 20%, so it seems like something was bothering him. Kudos to him adapting to changing defense by setting up Hyman to fill the net.
I don’t think Knoblauch is going to soon move away from McDavid/Hyman unless this carries on for a while.
For being a proclaimed math blog I finding it a bit ironic to be quoting stats with a 3 game sample. From a statistical point of view these numbers are completely meaningless. The margin of error in these statistics would be monstrous. The fact that there is no statistical error analysis in hockey analytics is fascinating and telling.
Now using these numbers to confirm the eye test, there they might have some use. But that is not math. Last I checked the UofA doesn’t have a statistics course on using statistics in conjuction with subjective interpretation.
There may not be a course on statistics in conjunction with subjective interpretation, but isn’t that the nature of hockey statistics?
I read sometime long ago that to eliminate the effect of luck in a hockey series, it would need to be a best of 77 (or something to that effect) because of the prevalence of random chance, luck and variability that exists in hockey. For comparisons sake, I believe baseball was a best of 15, and Basketball was a best of 7, and this is due largely in part to the rarity of goal-scoring events in hockey as compared to other sports.
This luck is part and parcel with hockey, and so any statistical analysis is largely useless without subjective interpretation. If a player passes the eye test and looks fantastic on the ice, it is worth it to double check statistics to verify their play. Lauri Korpikoski always looked to be moving on the ice, so did Ryan Jones, but a deeper look into their stats showed they didn’t do that much. On the other hand, players like Puljujarvi looked lost a lot of the time, but then would pop up with amazing Corsi, Fenwick etc… stats.
Analysis should not be limited to one or the other, but the eye test needs statistics, like statistics need the eye test. The math itself is largely useless, except to challenge bias and create a story of what has already happened.
What I see above is that the Oilers have outshot, and out-shot attempted their opponents, which feels like an indicator we have outplayed our opponents. The numbers themselves are an extremely tiny sample size, but its the only one we have right now. The eye test is more prevalent now, and statistics will be more prevalent later in the year when there is too much information to properly parse through.
I would love to see a thorough analysis of error in hockey, i think it would be fantastic. I also believe that the stats have value, even if they are not statistically relevant, even if they require subjective interpretation.
Statistics don’t need the eye test. The whole point of statistics is that it is a science and is the same regardless of the observer.
While we are on discussing on this blog which stats are relevant and which ones are irrelevant, there is an entire branch of mathematics dedicated to this exact question. A huge part of statistics is about separating signal from noise.
As an example say someone has a 65% DFF one game. But there is only a 30 percent chance of him repeating that the next game. What am I supposed to do with that statistic? That’s great the player had a good game but it doesn’t tell me anything about what he is going to do in the next game.
Also just to add to this. If a player has good statistics but isn’t helping the team win then it is the statistics that are bad. It’s all about correlation versus causation and it is desperately hard to separate them.
Remember when Dallas Eakins tried to game the system by coaching the players to improve their corsi? It failed spectacularly because good teams have good corsi but that doesn’t mean having good corsi makes you a good team.
In fact for players statistics are largely useless because it’s not about what they have done. It is about what they are going to do. It’s about making the right decision in the right moment. And that might be making the high percentage play or it might be making the low percentage play. It’s a case by case situation.
While it is a good sentiment, it requires a deeper analysis, and better statistics to allow an objective observer to make definitive conclusions without the eye test. As you said below, Dallas Eakins tried to game the system and failed because the stats by themselves aren’t worth a lot. A player taking 100 shots from behind his own net, would outperform a player with three shot attempts, even if he scored on all three attempts. The current statistics available to us are insufficient for your requirements. while you may be technically correct in that it is not ‘pure’ math, its still the use of statistics to support or negate prior bias, which in my opinion, is a fair use of math.
He literally used the numbers to confirm his eye test. The data is not meaningless because it’s small and he noted score effects that are skewing results.
Weird way to tell the world you’re smart.
This blog values math, it does not claim to be a math blog.
People call this a math blog regularly.
This is actually quite dismissive and rude. This is something I do know a bit about and have studied in university. I thought I would add to the discussion by sharing some of the knowledge I have learned but according to you that is just being arrogant I guess.
Statistics is an entire branch of mathematics with tons of theory and research. It’s not just divide this by that and draw conclusions.
I have seen the “small sample size” disclaimer so many times on this blog over the years from so many posters, including LT, that it always surprises me when this kind of comment pops up.
It would be like someone popping in one day to be like “hey, I think Chiarelli was not a very good GM, am I the only one?”
Oh I know people use the “small sample size” disclaimer all the time. Saying small sample size means that it is an untrustworthy statistic. In my opinion it just creates bias because adding a point that might be right or might be wrong just muddies the water.
Saying things like the Oilers are losing more games than they’re winning, so far this season, that isn’t a stat, it’s just early empirical data.
When you have little information you still run with it. When Germany invaded Belgium and France it’s not like people sat around and said “hmm, let’s wait for a few more years of data before we make any conclusions.”
The site isn’t about rigorous stats at all costs. It’s a hockey discussion site where people enjoy playing with stats to aid their insights.
Britain and France declared war on Germany PRIOR to the invasion of Belgium. That declaration was why Germany invaded Belgium. You need a better analogy. And if you think any of the parties was operating on minimal information you would be sorely mistaken.
You’d probably want 20-25 games if you’re using team-level stats to predict the rest of the season, but I don’t see anything wrong with using them to identify early trends. LT isn’t making any sweeping conclusions based on this data, just noting what the numbers show so far.
I know, right? Guy writes this crap every day! What’s the point? Blah blah blah music reference blah blah blah weird math blah blah blah burning daylight!
It’s a waste of time!! Why bother, right??
Stupid people are the worst. This Lowetide guy seems dull to me!!!
Time to pivot and turn this blog into a Ziggy blog.
Why are you mocking me? I never called you dull or stupid. Maybe my tone came off wrong but I don’t think I deserve that.
I don’t think I deserved your comment. I responded.
I don’t see anything insulting in his post that would justify mockery and the shutting down of the discussion. He has opened a can of worms and did so intelligently. The normal way to respond is with an intelligent defense, no?
I think it’s fine to have this discussion, and people can fill their hat.
I choose what I value. If you don’t like those things, post something you value. There is no need to rip what I’ve written.
I even SAID many people don’t value Corsi. I do. That’s the blog. I’m not going to change my view. You CAN express your view without insulting the blog.
You insist on respect and no name calling. Putting words in my mouth that I was calling you names feels like name calling in a roundabout way. You misrepresented me and that is disrespectful.
I think you might be quite interested in the mathematics behind statistics. I’m sorry if you were insulted by my this is not math comment.
I apologize to you for being unkind. I value this blog and do believe in the numbers I bring.
I went over to a friend’s house for a BBQ yesterday and made sure to let him know the burger was overcooked, the beer was warm, and, of course, I had to remind him that I went to Hamburger University, so I know a thing or two about cooking burgers I was just trying to add my expertise to the situation 😉.
I asked my uncle who was a stats prof at UBC about hockey and stats. He isn’t a team sports guy, but he said that stats should confirm what the eye sees. I’m not sure how that fits with what you are saying, that it isn’t math
When stats fail and don’t reflect reality in the end, which happens all of the time, isn’t it a matter of faulty interpretation? Data collection is the easier part. What to do with it isn’t, especially if the data set is big. Isn’t the interpretation always at least to some degree subjective?
Einstein didn’t come up with a math formula first, he started with a thought and observation
On top of that, even when you don’t have enough data to create a reliable stat, the data is still accumulating. It’s no crime to watch a reliable stat build up.
For example, on election night a lot of people watch the early returns and exit polls. You could say, that’s unreliable, you’re wasting your time! Especially early in the night.
But people still want to see the results form-up.
I would love to hear what your uncle has to say about the lack of error analysis in hockey statistics and analytics. I’m being honest. I studied math and physics but I only have a bachelor’s and I would love to hear from a more knowledgeable person.
My statistics classes were all about confidence intervals, standard of deviations and distributions. Which are things you never hear about in hockey analytics.
Einstein also refuted quantum mechanics because it was very abstract and based on mathematical anomalies. He also said later in life that it was one of his biggest mistakes.
The more variables involved and the more dynamic the reality on the ground, the more difficult any system is to analyze. Stats are famously impotent with regards to economics despite an entire field in economics devoted to statistical analysis of the economy. But an entire economy has so many moving parts all of which are changing in their impact and influence every second of every day that statistical analysis results in some startlingly opaque and often completely wrong conclusions.
That doesn’t stop them from trying. They invented a whole field of study to try. Hockey, out of all major team sports, has the most dynamic play confounded with the most variables. That doesn’t stop people from trying.
Your reticence to take a 3 game sample with much seriousness is healthy. But you can ask yourself, does what LT is describing statistically match your own viewing? If it does not you could point out where it does not, thus refuting the analysis.
We finally have a road game I hated the 4 home games to begin the year. This team needs to play a simple road game with the opposition pressing and the Oilers coming back the other way with odd man rushes.
I was not a fan of the Corey Perry resigning. I was wrong and must admit that he has provided more than some of the others so far!
Trolling through the Flyer’s players to find a willing combatant was gold. He has brought energy and truculence to a team badly in need of both. At 39 he may not be an everyday player – but if he can contribute either on the score sheet or physically (both?) for even 50 games then he is a valuable roster player.
Perry has brought better than 65% in both Corsi and Fenwick as well as significant positive relative values in both categories.
Perry is leading by example and needs to shame some of the others into playing a physical style of game.
Perry has a goal, a very questionable disallowed goal and 2 posts where he beat the goalie clean. Who knows if he can sustain it but right now re-signing him has been a stroke of genius.
I strongly suspect the full TC, lack of distractions and familiarity with his team are making a difference. He’s a prepare-and-be-ready type.
Lets not forget, Perry did have 12 goals in 54 games last year – that’s a, what, 17 goal pace?
Yup, agreed – Perry has been a good player for the Oilers so far this year.
I can still be (and am) miffed that he grinded for performance bonuses and is $375K over league min.
Prospectorono!
The Colts of Barrie are in the spotlight.
Wakely has come flying out of the gates in his overage season, putting up 3+2 in two GP with his new squadron.
Akey finally posted a crooked number, a goal last time out.
Puck drops @ 5 p.m. Hoadley time.
Programming note: TSN will be showing select CHL games again this season, including two games featuring NAmateurs in the coming weeks:
Friday, October 25 (London)
Friday, November 1 (Barrie)
Great intel on the TSN games – looking forward to those.
ON Wakely, here is hoping that he continue his high end production but, for me, I find it hard to ever put much stock in CHL numbers from an over-ager – i mean lets not forget Cam Hebig (and other that came before).
I expect close to 1.5 PPG for Wakely!