INTRODUCTION TO PDO
PDO By Line
1 – 100.3
2 – 100.2
3 – 99.7
4 – 98.6
PDO is on-ice save percentage + on-ice shooting percentage. The team at the top got the luckiest. Yes, they revert to 100 over time because its very unusual for an entire team to avoid being average (it’s a pro league).
Shooting and save percentages always normalize.
If you combine shooting and goaltending, you get regular PDO:
PDO = SH% + SV%
This is the format you would see on stats sites, but you could also calculate PDO based on fenwick or corsi:
FPDO = FSH% + FSV%
CPDO =CSH% + CSV%
The reason you look at PDO, is because shooting and goaltending are less predictable than generation and prevention (More on this in chapter 3). This means that there’s a higher variance in shooting and goaltending stats.
So, if the PDO is a lot higher or lower than 100%, there’s a good chance the results are unsustainable. You often hear that PDO will regress towards 100% over time, but this isn’t necessarily true – Some teams have better than average goaltenders and shooters.
Over time the PDO should regress towards the true quality of the shooters/goaltenders.
Earlier we discussed PDO. This was a way to combine goaltending and shooting into one metric. We can obviously do something similar with the more advanced goaltender/shooting stats. Here’s how I would define delta-PDO or dPDO:
dPDO = dFSH% + dFSV% + 1
dPDO can be used the same way regular PDO, but it’s a better metric since it includes shot quality.Alternatively, we could define analytics-PDO or aPDO as:
aPDO = GAx/xGF + GSAx/xGA + 1
This would be my preferred metric for combined goaltending and shooting.
TRIVIA CRAPS
(Aug 2008 – First Appearance)
I’m not saying that shooting % and save% aren’t important, they obviously are. Just that so much of it is pure shithouse luck, that sensible people who are trying to analyze this stuff are well advised to treat them as pure luck to start, and add an element in for them after.
WHAT IS PDO
(Nov 2008)
A PDO number is a team’s ES S% plus their ESSV%.
there’s a lack of sustainability in that number – if it’s high, it’s going to come down; if it’s low, it’s going to go back up.
Not just teams, either: players too.
IF YOU WERE TO UNDERSTAND ONE STAT
1. Shooting percentage is primarily luck-driven
2. Save percentage is primarily luck-driven
3. The apparent talent there is in these quantities, particularly offensive finishing ability, shows up as negative talent at the other end of the ice
PDO REGRESSION TO MEAN
Future PDO regresses to the mean relative to PDO though a certain number of shots (both for and against.) For example, PDO through the first 1000 shots regresses 87% to the mean over the rest of the season – a team with a 1030 PDO through 1000 shots expects a PDO of 1004 for the rest of the season.
PDO STREAKS & SEASON LONG PERFORMANCE
If your team is posting a 12% shooting percentage or a 1040 PDO through 10 or 20 or 30 games, you need to ask yourself if it’s illusory, or if your team is the best team we’ve seen in half a decade.
The spread over a full season is roughly half as large as the spread over 20 games. So whatever you observe in a quarter-season needs to be tempered by 50%; and don’t be surprised if a team riding a PDO high starts losing as they come back to earth.
THOUGHTS ON PDO & LUCK
(Feb 2012)
not all players have a natural PDO level of 1000 and there are two reasons why.
1. Not all players play in front of perfectly average goalies which will have a major impact on the save percentage portion of PDO.
2. Players can drive shooting percentages.
PDO varies from 983.5 up to 1012.4 depending on the group’s ice time. This is largely driven by shooting percentage which varies from 7.5% to 10.4% with the players with the lowest amount of ice time having the lowest on-ice shooting percentage and the players with the most ice time having the highest shooting percentage.
Driving on-ice shooting percentage is a skill. This means more talented players can have a natural PDO (the PDO that they should regress to) above 1000 and less talented players can have a nautral PDO below 1000. Factor in the goaltending and a player could have a natural PDO well above or well below 1000.
Definitionally, league wide it must average to 1000 but that doesn’t indicate a reason that for a given player, line combination or team it should
PDO AND LUCK
(May 2012)
The general idea, though, is that the majority of league talent will float around some kind of average. Finding the disparity between a player’s output and the league average concerning luck, though, is another difficult task.
The statistic operates on three principles.
(a) Shooting percentage is primarily luck-driven
(b) Save percentage is primarily luck-driven
(c) Whatever apparent talent there is in these quantities, particularly offensive finishing ability, shows up as negative talent at the other end of the ice
for the vast majority of NHL regulars, a high PDO in one season comes crashing down the next. And many players with high on-ice shooting percentages get them by cheating offensively, which leaves them susceptible to higher-percentage opportunities against them at the other end of the ice.
INTRO TO PDO
Teams can consistently outperform PDO as a result of their excellent team goaltending and shooting.
Median SH% over the past 5 years in the NHL is 8.3% at 5v5 while SV% is .917.
At the team level PDO will allow you to quickly spot a team unlikely to sustain its performance mid-season.
At the team level, slightly over half of team talent exhibits itself as puck possession, and slightly under half exhibits itself as finishing (and preventing finishing). It just so happens that the possession talent is much, much easier to measure.”
Look for a disconnect between possession (CF%/FF%) & PDO.
it’s extremely difficult for a forward to influence On-Ice SV%
when we examine PDO for players, it is not assumed that the PDO score should regress towards 1000 as is often suggested.
Again, look for a disconnect between possession (CF%/FF%) & PDO.
STUDYING LUCK & OTHER FACTORS
no one has clearly demonstrated how large a deviation from the mean is to be expected, over specific samples of games.
a team’s PDO tightens closer to their true mean (1 for a league average team) as teams play more games.
Over an 82 game season both the “talent” and “random” parts of PDO accounts for about ± 6 points.
The graph above may be the most important and most useful graph of everything presented in this study. The graph represents the random variation expected over a sample of games for PDO (See Methods above for model details) indicating 3 standard deviations in red. The gray area represents greater than 2 SD. This is the “danger” area in which a PDO would be unlikely to be strictly due to chance, although still possible. Essentially the graph represents the the amount of “give” we might suspect for a team’s PDO (y-axis) over a sample of games (x-axis).
As you see, this graph is very valuable for analyzing a team’s PDO. We can use the lines as a test of statistical significance. Any team with a PDO that differs from 1.0 and enters into the gray zone, (eg. A team with a PDO greater than 1.048 or less than 0.952 at 10 games and so on) would lead us to believe that their PDO is a result of something that team is doing, and probably not normal variation.
The more stringent red line represents 3 SD, which leads us to believe that a given team’s PDO is very likely to be due to poor or great play; as only 1 in 500 teams are outside of these values (>1.07 or <0.92 for 10 games) by random fluctuation alone.
2.) Factors That Influence PDO
The largest drivers are not unexpected, Shot% representing the bulk of movement, power-play differential, and home/road.
PDO has a dramatic effect on the likelihood of a team winning.
PDO will completely determine the outcome of the game.
· PDO regresses heavily to the mean with a sharp decline in the standard deviation as the number of games increases.
· The factors that may influence PDO were home/road, PPdiff (power-plays for minus power-plays against), PtDiff (year-end points – year end opponent points, excluding points accrued from that match), Shot% (Shots For – Shots Against) were mostly insignificant, but were statically measureable, in total explaining 12% of the change in PDO
· Lastly, we can calculate how an average team would perform, given a specific PDO using table 5 (see below). This gives us an understanding of how a team is performing independent of their PDO.
· Talent and random portions of 1 SD of PDO both account for approximately ± 6 points at the end of the season.
· At about 70 games, the talent portion of PDO equals the random portion
LIES, DAMNED LIES, AND PDO
While PDO may be a good indicator of team luck, it’s not necessarily a good indicator of individual player luck.
PDO is meant to measure how much of a team’s goal differential is due to random variance, rather than more demonstrably sustainable abilities like puck possession, and for the most part, others have shown it does that quite nicely.
Elite players demonstrate consistently superior shooting abilities. Shooting% of first liners do exceed that of fourth liners by quite a significant margin
1.000+ PDOs for first line forwards are not due simply to good luck. Similarly, sub-1.000 PDOs for third and fourth line players are not necessarily due to poor luck either.
1) PDO does not capture luck on an individual skater level.
2) The mean that PDO can be expected to regress to is unique to each player, and is influenced mainly by that players role and shooting ability
I don’t think using PDO to make judgments on an individual player’s point production is particularly wise.
PDO is good for the broad-brush analysis on a macro-scale (“the Ducks’ record is unsustainable!”), but it’s “predictive” power doesn’t really carry over to a micro-scale, because you’re not dealing with nice, round numbers anymore and PDO’s two components are completely unrelated to each other.
WHY STATS GUYS CONFLATE PDO & LUCK
(April 2014)
I can basically choose any cutoff that I want – 10 games, 60 games, two years – and identify the players with high and low PDO. From that point forward, my best guess for the average PDO of the group is basically 100.
Your best guess going forward for groups of players who have extreme PDOs is generally right around 100.
It doesn’t tell us much about the future. So it’s luck.
CAN TEAM SYSTEMS INFLUENCE PDO
Still, insofar as they depend on good fortune, high PDO values are generally not viewed as predictors of sustainable success.
Still, this thinking suggests that the only meaningful sources of variation in PDO are player talent and puck luck, leaving a very important question unanswered: are there features of team systems that can contribute to lower or higher PDO?
The obvious implication is that high shot volumes are associated with lower average shot quality and higher save percentages. To the extent that a team’s style is predicated on maximizing shot creation and/or minimizing shots against, it would seem that this style would end up partially determining their PDO.
On average, positive possession teams have a PDO of 1.000, while negative possession teams have an average expected PDO of 1.001. Thus, it appears that strong possession hockey has a slightly negative relationship with PDO.
On average, high-event teams have a PDO of 1.002, while low-event teams have a PDO of 1.001. Possibly owing a lot to the save-percentage effect, firewagon hockey appears to be slightly associated with higher PDO.
When you bring the math of possession into things, trading off possession for PDO starts to look like less of a brilliant idea. For a team playing very low-event hockey, there could be an argument for loosening things up slightly to take advantage of the Sh% effect associated with average rates of shot creation. If you’re a decent possession team looking to take advantage of this effect, your priority should be on at least maintaining your possession differential;
FORMULATING A NEW STATISTIC FOR LUCK
while team shooting percentage does seem to be random over time, team save percentage shows distinct evidence of not being random.
The influence from SV% was so strong that it made PDO statistically predictable over time — kind of a bad thing when it’s traditionally used to reflect luck.
we need to find an expected save percentage where it is EQUALLY LIKELY for his actual SV% to be above or below that number.
I cannot accept a hypothesis that one year’s MPDO has any bearing on the next year’s — the relationship is random.
Luck in hockey is possible. Showing which teams are riding luck and which ones are due to break-out has been one of the most useful developments of the advanced stats community.
However, I think I’ve done a pretty thorough job of statistically proving that traditional PDO is not a random measure. That’s what this entire thing is about — if PDO is not random, than it does not reflect luck.
TML & SHOT QUALITY
teams rarely sustain good records without a decent shot differential.
in today’s NHL, the differences between teams in shooting percentages are small. Over a month or two, teams can run hot enough to beat their shot differential, but in recent memory those runs have all proven to be transient.
shot differential drives their results more than shooting percentages do.
· The vast majority of the NHL spends the vast majority of its time in that middle band from about 9-10 percent.
· The league average appears to have drifted down a bit (mostly the result of a reduction in power plays).
· Toronto’s run isn’t unprecedented. About once per year, a team will pop up with a >12 percent shooting percentage over a 50-game span.
· Sustaining such a run would be unprecedented in recent history. An appearance above 12 percent is generally a brief spike, and the longest stretch of maintaining an average above 11 percent is 67 games.
The best any of Toronto’s predecessors did over their next 50 games was 10.7 percent.
PDO is an indicator of how much a team has outperformed (or underperformed) its shot differential.
If the Bruins – with record-setting goaltending and one of the deepest forward groups in the game – can’t sustain a 102.0 PDO,
EXPECTED PDO – HOW SUSTAINABLE IS BEING OUTSHOT
In a sense, the team with the better SV% can tolerate more shots against than the average team with average goaltending. Further, a team with better team SH% can tolerate fewer shots for than the average shooting team.
The chart below is the PDO required to win given various SF% (shot differentials). That is, these PDOs on average will give an all situation positive goal differential and thus winning record before shootouts for the given shot differential.
40% - 1028
42% - 1024
44% - 1019
46% - 1013
48% - 1007
50% - 1000
51% - 997
Another method to find “true PDO” based on skill components is to create an expected PDO for the team based on the career average SH% and SV% of each individual player/goalie.
For example, a low SH% team like NJD or the Sabres would be expected to have a below average expected SH% and thus low PDO.
SHOT DIFFERENTIAL & PDO
How important is shot quality – is it almost as important as shot differential, almost completely inconsequential, or something in between?
As an explanatory variable, it [shot quality] is 4 to 5 times less important than shot differential.
It seems that no matter how you approach the problem, everything converges on something in this vicinity. Shot quality isn’t insignificant, but it’s clearly subordinate.
Still, we often use PDO as a measure of luck in the short term. It’s true that we don’t necessarily expect everyone to be right at 100 in the long run, but when we use PDO to say that a team has gotten lucky, it’s generally because they’re way above 100. When we’re talking about Toronto at 104 or Colorado at 106, we can safely project significant regression without worrying about whether the team will be 99.6 or 100.7 in the long run.
By far the most common use of PDO is as a first screen to pick out the teams running so hot (or cold) that it would be crazy to expect them to continue at that level.
Basing a projection on a team’s shooting and save percentages over the short run – with all the attendant issues of randomness – is a very shaky prospect.
OUTPERFORMING PDO
The actual spread of PDO, from roughly 975 to roughly 1025.
AvgPDO = 1000 +/- 25
we might use all-situations PDO to see if there are teams that either exhibit consistently higher/lower-than-average PDO (due to talent) or because they artificially affect it with more/less powerplay than shorthanded time.
The actual distribution of ST% is quite shallow
Which means special teams time overall doesn’t tip very far (nor should be expected to tip very far) in either direction.
the shallow distribution and luck in those percentages nearly eliminates Special Teams Percentage (ST%) as an effect on all-situations PDO.
So, if special teams time isn’t expected to affect PDO, could a team still exhibit PDO “talent?”
you can see the strength of the pull of PDO’s 1000 average. It’s so strong that it really keeps us from labeling any more than maybe 5-7% of the last 300 team performances (just under one team a year) as indicative of anything that might be called PDO “talent.”
DANGERS OF SH% & SV% GRAPHS
Regression is a dangerous word. Teams don’t all regress to the same values, or at the same rate. Basically, tread with caution.
What the percentages chart does is punish teams for scoring and preventing excess goals. It’s basically saying, “If you score or prevent more goals than you need to win, your winning is less sustainable,” which we in fact know to be the opposite of the case.
Shot quality matters, but we know it matters far less than shot rates, and therefore — at least early in the season — shot rates are the best quick way to examine a team’s start
GOAL METRICS BETTER THAN SHOT METRICS
(Aug 13, 2016)
If shooting and save % are in fact luck based, PDO% should average about 100 in the long term, for every single team.
Statistically we expect luck to balance out over time,
This does not appear to be the case when it comes to shooting and save%, as PDO did not even out over the 10 season span
fluctuations in PDO are the results of some teams taking better shots, some allowing better shots, and some teams simply having better goaltending.
HDSC & PDO
when you hear people call PDO the “luck” statistic, that’s not entirely true – it is partially a product of a team having disproportionately more higher danger chances than their opponents which are more likely to lead to goals.
Players who receive more ice time have a higher PDO.
PDO doesn’t regress to 100 at all times: Over time it will always revert to the natural PDO,
I believe natural PDO depends on 3 things:
· Team strategy – If you value quality over quantity (think Trotz style hockey)
· Goaltending – A good goaltender leads to a higher save percentage
· Shooting – Good shooters and shot distribution matters. You want your best shooters to take as many shots as possible.
PERCENTAGE LUCK IN HOCKEY EXPLAINED
(Nov 2020)
There are elements of PDO that aren’t caused by luck, like shooting talent and goaltending.
Some teams are just better at scoring and stopping pucks than others. But these are extreme examples, and overall repeatability of PDO is still extremely low.
PDO regresses to the team’s natural mean, not necessarily the league’s.
To conclude, the notion that full-season team percentage results are absolutely based on talent and not luck is untrue.
Aside: 11% on-ice SH% is extremely high for a player
The probability a player with similar results nearing, let alone sustaining, high on-ice shooting is quite low.
Comments