Jump to content

Let's talk PDO [1/2]


Quik

Recommended Posts

Was going to make this a short VHL.com, but it kinda grew, so posting it now since the stats won't really match once todays sims are uploaded

 

A recent(ish) trend in the hockey community is to look at predictors of future performance, rather than actual results that have occurred. Everyone knows that actual results can sometimes have random events that pollute the data. A cheap goal here, a goalie saving a puck they had no business saving there, and a team loses 4-1 in a game they dominated and probably should have won.

 

PDO is one of the most basic predictors of luck, in hockey. It is Team-wide Shooting % + Team-wide Save %. 100.00 is the expected average. Anything higher, and a team is running on a bit of luck, while anything lower, and the team has been unlucky. League average so far this season is 99.93%, so pretty much exactly what you'd expect. For a bit of a primer, last season the average PDO was 99.96, and no team was more than 0.79 away in either direction, which indicates that all teams will regress to the mean, eventually.

 

When looking at individual teams, this season, there are 4 distinct categories you could divvy up the VHL's 9 teams into: Very Lucky / Average / Unlucky / Very Unlucky.

 

We'll start with the teams who have been average, in terms of PDO, and thus playing at about the level they would be expected to play over the course of their whole season. In the order they sit in the standings, they are Vancouver (100.25), Davos (99.86) and Moscow (99.87); all within 0.25 of the 100.00 mark. In broader terms, it means none of these teams are more than a 2-goal differential away from where they would expect to be, and should probably continue to play at their respective paces for the remainder of the season (without accounting for streaks during the season).

 

The next closest group to "average" is the "unlucky" category. These are teams that have below average PDOs, but are still within 1% if league average. These two teams that are slightly under-performing are Riga (99.11) and Calgary (99.04). If both these two teams had been pacing at league average S% and SV%, they would both be expected to have an extra 5 goals in their differential (Riga would be +22, rather than the +17 they currently sit at, for example). Over the course of a 72-game season, you could expect both the Reign and Wranglers to improve a fair amount, as they regress up towards the league average.

 

After the "unlucky" group, the next closest group to "League Average" is the Very Unlucky group. This group contains a pair of teams that are at least 1.35 off the median PDO, and one would expect significant improvement from as the season wears on. They New York (98.64) and Helsinki (98.41). Both teams are currently 11 goals below what their expected goal differential would be, and could be expected to go on winning streaks at any moment. Helsinki, in particular, is one of two teams to have played all 9 of their backup goalie starts, and is probably expected to see some significant upward regression!

 

As for the final, "Very Lucky" group, that includes the two teams currently occupying 1st and 2nd place in the league standings. There is something to be said for talent outplaying PDO, as the theory that better players will perform better is a prevalent one. However, both teams are currently playing more than 2% above league average, and will probably see some downward regression as the season continues. The Seattle Bears, sitting first place in the league, hold a 102.00 PDO, playing to a 17 goal-differential better than their S/SA would indicate. The Legion, meanwhile, currently hold an even more absurd 102.21 PDO, outpacing their expected goal-differential by 18 goals!

 

While the disclaimer that the season still needs to play out, and things could change at any moment, one would expect things to level out as the season wears on. Based on predictors like SF and SA, it is highly likely that the two teams at the top will see a fall back to the pack, perhaps even the middle of it, while Riga should be expected to rise to the top, and New York and Helsinki continue to battle for the final two playoff spots with the likes of Toronto and Vancouver.

 

Current Season PDO (top chart) vs S64 PDO (bottom chart)

pdo.png

Looking at that spreadsheet, a positive 'DiffVSxDiff' would mean a team is outperforming their indicators, while the opposite would also be true for a negative 'DiffVSxDiff'.

Link to comment
Share on other sites

24 minutes ago, flyersfan1453 said:

@Quik, how did you come up with xG and xGA? I thought those stats took into consideration shot quality, which we don't have access to.

It was more a rudimentary one based off SF/SA against average S%/SV%. 

Link to comment
Share on other sites

13 hours ago, Spade18 said:

so what youre saying.  Is I can blame our loses on having bad luck.

 

That sounds about right.

 

That's easier to accept than the real reason 

 

<---------------

Link to comment
Share on other sites

4 hours ago, BOOM™ said:

 

That's easier to accept than the real reason 

 

<---------------

Is it your post count? Helsinki didn't have a problem winning with Vaydar (Last 4 games notwithstanding...)

Link to comment
Share on other sites

I haven't checked, but is Roctrion a better build than Odinsson? I always have higher expectations and talk my players down if they are disappointing to me. I expect better from my fake hockey creation. 

Link to comment
Share on other sites

7 minutes ago, BOOM™ said:

I haven't checked, but is Roctrion a better build than Odinsson? I always have higher expectations and talk my players down if they are disappointing to me. I expect better from my fake hockey creation. 

 

Fairly similar, except you've got higher passing and checking. It's probably my GMing that brought your stats (and now his :() down

Link to comment
Share on other sites

5 hours ago, Quik said:

 

Fairly similar, except you've got higher passing and checking. It's probably my GMing that brought your stats (and now his :() down

 

Doubtful. I blame that Simon guy! 

Link to comment
Share on other sites

On 3/15/2019 at 2:03 PM, BOOM™ said:

 

Doubtful. I blame that Simon guy! 

 

Aye, I'd say so too. Honestly the biggest question mark of the last few seasons Stopko notwithstanding has been the goaltending in the VHL. From Cast getting that crazy successful season while goalies suffered, to inconsistencies last season and now even wilder inconsistencies this season, I don't know what it is. Goalie stats should have averaged higher than they are, and the idea that we could somehow see sub .920's for most...yet again? Goalies of that TPE levels used to have above .925 most of the time, if not .930's.  It's making me start to wonder if the harshness of our update scale change perhaps had a bigger effect on netminders than we thought?

 

Yes Stopko was a god, but didn't he like nearly break the TPE record? A goalie shouldn't need 1000 TPE to be good. 

Link to comment
Share on other sites

  • Commissioner
11 hours ago, Devise said:

Yes Stopko was a god, but didn't he like nearly break the TPE record? A goalie shouldn't need 1000 TPE to be good. 

Honestly too looking at his regular season stats, was he a god? He played in a tougher time sure and I agree on god status in the playoffs but regular season he doesn’t come close to touching the goalies of the 50’s.

 

My guess is more and more skaters have figured out the formula to play well which results in more goals being spread out instead of massing on a few players. This means inconsistent forward play but in general bad goalie play because there’s no real “secret sauce” for goalies.

Link to comment
Share on other sites

12 hours ago, Beketov said:

Honestly too looking at his regular season stats, was he a god? He played in a tougher time sure and I agree on god status in the playoffs but regular season he doesn’t come close to touching the goalies of the 50’s.

 

My guess is more and more skaters have figured out the formula to play well which results in more goals being spread out instead of massing on a few players. This means inconsistent forward play but in general bad goalie play because there’s no real “secret sauce” for goalies.

 

Possibly a related thing, since more skaters are opting for the highest offensive output, it's causing their defensive stats to be worse by proxy (especially since you can't really see defensive metrics in the sim very well). So you have better offenses playing against worse defenses which makes it harder on the goalies.

Link to comment
Share on other sites

  • Commissioner
Just now, omgitshim said:

 

Possibly a related thing, since more skaters are opting for the highest offensive output, it's causing their defensive stats to be worse by proxy (especially since you can't really see defensive metrics in the sim very well). So you have better offenses playing against worse defenses which makes it harder on the goalies.

Yeah, that could be the case as well.

 

 

Link to comment
Share on other sites

pdo3.png

 

  • Toronto continues to fly above 102, riding a shooting % that's nearly 2% above the league avg.
  • Seattle continues to flirt with 102 as well
  • Moscow dropping like flies, going from 99.57 to now 98.21
  • Helsinki and New York both see a slight uptick, but nowhere near where they should be
  • Based on xDiff (still not a perfect measure), teams would be arranged as:
    1. Seattle (+29, 131 xGF)
    2. Riga (+29, 130 xGF)
    3. Helsinki (+18, 126 xGF)
    4. Toronto (+13, 123 xGF)
    5. Vancouver (+5, 112 xGF)
    6. New York (+4, 118 xGF)
    7. Davos (-16, 106 xGF)
    8. Calgary (-26, 94 xGF)
    9. Moscow (-59, 91 xGF)
Link to comment
Share on other sites

  • 2 months later...

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...