Calgarypuck Advanced Stat Summary

[edslunch]

Originally Posted by Dan02

OT isn't, however the shootout is. The stats have been run and as long as you put your best shooters out there all the time, over the long run the shootout is essentially a coin flip.

Fair enough, but the Flames are only 3-1 in the shootout, thus a remarkable 7-2 in 4 on 4. The Flames proficiency at 4-4 - a situation that suits their style of play - isn't measured by 5v5 Corsi. Nor is their stellar penalty performance - while they are neither very good at PK nor great at PP, the differential in number of penalties taken gives them a significant special teams advantage.

[Loyal and True]

Originally Posted by edslunch

Fair enough, but the Flames are only 3-1 in the shootout, thus a remarkable 7-2 in 4 on 4. The Flames proficiency at 4-4 - a situation that suits their style of play - isn't measured by 5v5 Corsi. Nor is their stellar penalty performance - while they are neither very good at PK nor great at PP, the differential in number of penalties taken gives them a significant special teams advantage.

A thousand thanks for your observations about 4 on 4 and penalty differential.

http://www.nhl.com/ice/teamstats.htm...enaltyKillTime

Flames have almost 106 minutes of extra 5 on 4 time because they don't take many penalties 5 on 5 (which may partially explain why teams have better 5 on 5 possession stats). 5 on 5 defensively , the Flames focus on inside position, using sticks in lanes and blocking shots. It's almost like a penalty kill in their own end, and it works.

They have 103 extra shots on PPs vs opponents' PPs. Teams usually shoot 13% on PP vs 8% 5 on 5. That massive penalty differential should result in 13 goals differential.

Not only that, Flames PK save percentage is around 28th in the league. Opponents have been getting lucky bounces and should have scored fewer goals on PK. Special teams PDO is around 95 or 96 when it should be 100 (according to PDO guys).

So yeah, they have high shooting % 5 on 5, but their PP and PK shooting and save % is due for big improvement. If those percentages even out, Flames could still be a playoff calibre team.

Most analytics guys overemphasize the 5 on 5 part of the game, while a third of the game is played either special teams or 4 on 4. In the Flames case, their PP time vs PK time is massive and that should offset the 5 on 5 corsi etc.

[Finger Cookin]

Originally Posted by edslunch

Fair enough, but the Flames are only 3-1 in the shootout, thus a remarkable 7-2 in 4 on 4. The Flames proficiency at 4-4 - a situation that suits their style of play - isn't measured by 5v5 Corsi. Nor is their stellar penalty performance - while they are neither very good at PK nor great at PP, the differential in number of penalties taken gives them a significant special teams advantage.

All Situations (3640:56 TOI)
GF% - 52.6      FF% - 48.0      CF% - 47.4      PDO - 101.43

5v5 (2709:29 TOI)
GF% - 49.5      FF% - 45.5      CF% - 44.6      PDO - 101.12

5v5 Close (1655:00 TOI)
GF% - 42.7      FF% - 45.3      CF% - 44.8      PDO - 98.87

4v4 (94:36 TOI)
GF% - 69.2      FF% - 49.6      CF% - 46.9      PDO - 110.64

The 4v4 results are definitely not explained away completely by luck or randomness, but there are definitely some eye popping differences. Compared to basic 5v5, the CF is ~5% better, and the FF is ~10% better - but PDO is almost 10% better (thanks to a more than 100% increase in shoooting %) and the GF% is almost 40% better. Expecting to maintain that kind of success seems far fetched, regardless of how well the Flames style of play is suited to regular season OT.

With about 1/3 of the season left, the Flames can't expect to go 5-1 in OT/shootouts down the stretch because of our past results. I still will cheer for the Flames to somehow do so.

[Finger Cookin]

Originally Posted by Loyal and True

Special teams PDO is around 95 or 96 when it should be 100 (according to PDO guys).

I don't think any analyst would expect special teams PDO to be 100. 5v5 situations, sure - but not special teams. One team has a distinct advantage, one team is at a distinct disadvantage. PDO would skew accordingly.

[Dan02]

Originally Posted by Finger Cookin

I don't think any analyst would expect special teams PDO to be 100. 5v5 situations, sure - but not special teams. One team has a distinct advantage, one team is at a distinct disadvantage. PDO would skew accordingly.

I think by that he meant on average the sum of your pp% and pk% should be 100% using pdo logic

[Loyal and True]

Originally Posted by Finger Cookin

I don't think any analyst would expect special teams PDO to be 100. 5v5 situations, sure - but not special teams. One team has a distinct advantage, one team is at a distinct disadvantage. PDO would skew accordingly.

Power play shooting percentage 13% plus PK save percentage 87% = 100

on PP your PDO should be about 105
on PK your PDO should be about 95

[Loyal and True]

Originally Posted by Finger Cookin

All Situations (3640:56 TOI)
GF% - 52.6      FF% - 48.0      CF% - 47.4      PDO - 101.43

Since we have massive differential of PP time (PDO 105) vs PK time (PDO 95) we would be very unlucky to have All Situations PDO of 100.00

[Finger Cookin]

Originally Posted by Loyal and True

Power play shooting percentage 13% plus PK save percentage 87% = 100

on PP your PDO should be about 105
on PK your PDO should be about 95

Summing up the last 7 seasons data for all teams across every PP and PK situation, PP PDO is 103.95 and PK is 96.05. So +-4, not 5. Pretty close though.

[Loyal and True]

Originally Posted by Finger Cookin

Summing up the last 7 seasons data for all teams across every PP and PK situation, PP PDO is 103.95 and OK is 96.05. So +-4, not 5. Pretty close though.

104 and 96 sounds good to me. Thanks for looking into it

[Enoch Root]

Originally Posted by Loyal and True

Since we have massive differential of PP time (PDO 105) vs PK time (PDO 95) we would be very unlucky to have All Situations PDO of 100.00

Strictly based on statistical expectation, the extra PP time only increases the expected PDO from 100.00 to 100.12

(this ignores quality of the PP and PK, etc. - just the TOI and an expected PDO of 104 and 96)

Edit: that may be flawed however, as PP and PK time might be more important than 5 on 5 time (i.e. more shots and more goals generated per minute). If so, then the impact would be greater than 0.12

[Street Pharmacist]

Originally Posted by Enoch Root

Strictly based on statistical expectation, the extra PP time only increases the expected PDO from 100.00 to 100.12

(this ignores quality of the PP and PK, etc. - just the TOI and an expected PDO of 104 and 96)

Edit: that may be flawed however, as PP and PK time might be more important than 5 on 5 time (i.e. more shots and more goals generated per minute). If so, then the impact would be greater than 0.12

Best way to do it is an awful lot of addition and subtraction.

[Enoch Root]

A better way to look at it is to say that their expected PDO on special teams is 100.79

They have spent 537.3 minutes on special teams, vs 2822.7 minutes of even strength (ignoring OT time)

So the next question is: how impactful are special teams vs even strength? (I don't have an answer for that, though a good place to start might be the percentage of goals that are scored on special teams)

[edslunch]

Originally Posted by Finger Cookin

With about 1/3 of the season left, the Flames can't expect to go 5-1 in OT/shootouts down the stretch because of our past results.

Apparently they can

At some point this gets far away from statistical probability that we have to assume there's something else going on

[Resolute 14]

Originally Posted by Loyal and True

Most analytics guys overemphasize the 5 on 5 part of the game, while a third of the game is played either special teams or 4 on 4. In the Flames case, their PP time vs PK time is massive and that should offset the 5 on 5 corsi etc.

FWIW, the Flames are closer to 80% 5 on 5 this year. The analytics guys focus on 5 on 5 play because situationally, it is most predictive of overall success. The Flames are actually an example of that. Our PK is horrible, and our PP has had long stretches of awfulness - after it started out so hot. So we've stunk at both ends of special teams, and our PP has shown that you can't really base any predictions off of it. That really just leaves even strength as our benchmark.

[edslunch]

Originally Posted by Resolute 14

FWIW, the Flames are closer to 80% 5 on 5 this year. The analytics guys focus on 5 on 5 play because situationally, it is most predictive of overall success. The Flames are actually an example of that. Our PK is horrible, and our PP has had long stretches of awfulness - after it started out so hot. So we've stunk at both ends of special teams, and our PP has shown that you can't really base any predictions off of it. That really just leaves even strength as our benchmark.

Yet aren't special teams (goal differential not effectiveness) and 4 on 4/OT the main drivers of their success?

[Resolute 14]

Originally Posted by edslunch

Yet aren't special teams (goal differential not effectiveness) and 4 on 4/OT the main drivers of their success?

I would say no.

We don't take many penalties, which limits the damage on special teams goal differential, and that is a function of our five on five play. And, as I noted, our power play has been wildly inconsistent.

Our 4 on 4 play is a driver for success, but we have played only 101 minutes at 4 on 4 vs. over 2700 minutes at even strength. We don't even get to those overtime games without strong 5 on 5 play. Or more accurately, without our huge shooting percentage at 5 on 5.

[Bandwagon In Flames]

Here's a current representation of Advanced Stats in regards to the Flames:


Image courtesy of Homer J Simpson from HFboards.

[Street Pharmacist]

Originally Posted by edslunch

Yet aren't special teams (goal differential not effectiveness) and 4 on 4/OT the main drivers of their success?

They partly explain it, but 5v5 is by a far the driver for every team. (as Resolute said)

Worth noting, the Oilers have a much better record since Todd Nelson behind the bench. You see all kinds of analysis from fans and media about how he's a better coach, how he deploys the troops, yada, yada. Here's something to look at:

5v5 goals % (percentage of the goals in their games scored by the oilers)

Before Nelson: 39.0%
After Nelson: 40.7%

Five in five they're the exact same horrible team. He's only improved special teams

[edslunch]

Originally Posted by Street Pharmacist

They partly explain it, but 5v5 is by a far the driver for every team. (as Resolute said)

Worth noting, the Oilers have a much better record since Todd Nelson behind the bench. You see all kinds of analysis from fans and media about how he's a better coach, how he deploys the troops, yada, yada. Here's something to look at:

5v5 goals % (percentage of the goals in their games scored by the oilers)

Before Nelson: 39.0%
After Nelson: 40.7%

Five in five they're the exact same horrible team. He's only improved special teams

I'm confused. You and Resolute make the point that 5v5 is by far the driver for every team, yet 5v5 stats don't line up with the Flames' (or the Oilers' recent) success. Do we conclude that they are outliers doomed to fall back or do we look at something else to explain it?

[Street Pharmacist]

Originally Posted by edslunch

I'm confused. You and Resolute make the point that 5v5 is by far the driver for every team, yet 5v5 stats don't line up with the Flames' (or the Oilers' recent) success. Do we conclude that they are outliers doomed to fall back or do we look at something else to explain it?

No, you're missing the point. Their 5v5 is driving the team, but their 5v5 is bucking the corsi expectations. It's still 5v5 that is driving success