Early Season Numbers: Which Flyers Will Revert To The Mean

We're only nine games into the season, and already fans and media personalities are beginning to draw conclusions and make evaluations, specifically regarding Mike Richards and Jeff Carter. Richards scored his first goal of the year on Saturday, while Carter is strangely dealing with trade rumors and contract negotiations at the same time. It's no surprise that Carter and Richards receive the most attention - the Captain and the goal scorer - and it's also no surprise that conclusions are drawn through nine games.

And honestly, there's no denying that Richards' one goal in nine games is pretty disappointing. At this rate, he'll only score nine goals all year. In addition, Carter has been subject to trade rumors for at least the past three years, so that's nothing new either. The problem is that proper evaluations are extremely difficult through nine games, largely due to a small sample size.

As an example of this small sample size, Blair Betts is on pace to score 18 goals this year, exactly twice as many as he's ever scored in a season before. Danny Briere, meanwhile, is on pace to score more goals than anybody in the NHL last year - 54 goals - or 22 goals more than he's ever scored before.

So let's look at three examples of players who will improve their numbers and three who will see their numbers fall. In other words, who will regress toward the mean?

Nowhere to Go But Up

1. Mike Richards' plus/minus

If you've read any of my previous articles, you'll know how much I hate plus/minus. In fact, I've been meaning to write about that issue specifically using data from previous years, but it will have to wait. Instead, we'll look at it in the case of Richards. So far, Richards is a minus-2 on the year. Meanwhile, Richards has been on the ice for seven goals against and five goals for five-on-five. We can draw many conclusions from this - he was playing top talent, he was playing with poor teammates, he was doing his job, but his teammates weren't, he was defensively responsible for those goals, etc. - but that's not the point.

Instead, Richards' plus/minus will improve simply because of poor luck. Per 60 minutes of even-strength play, Flyers goaltenders only see 23.6 shots while Richards is on the ice, good for 4th fewest among forwards. Despite this, Flyers netminders are only stopping 86.0% of those shots, leaving Richards with the worst goaltending on the team. Since Flyers goaltenders are stopping roughly 91.0% of shots at five-on-five, Richards will surely be on the ice for fewer goals against going forward.

2. Jeff Carter's shooting percentage

Currently, Carter is on pace to score 27 goals. This would be his lowest total in the last three years, and is definitely below what is reasonably expected of him. So far, he has an 8.3% shooting percentage (excluding the percentage of shots he fires that miss the net) which would also be his lowest in three years. Even last year, he registered a 10.3% shooting percentage, the second lowest of his career.

Obviously, shooting percentage is very unstable, but it is unlikely that Carter stays over 2.5 percentage points below his career level. Compounding matters - and relating to Carter's plus/minus - the Flyers as a team only have a 5.97% shooting percentage while Carter is on the ice. Just like with Richards, luck will begin trending toward Carter and he'll start to see the puck behind the opposition goalie more often, both off his stick and his linemates'.

3. Sergei Bobrovsky's save percentage

I know, I know: this is not by any means a guarantee. Because he's a rookie, there is very little track record by which to judge him on, but the numbers suggest - and I'm sure your eyes agree with you - that his 0.904 save percentage is lower than it should be. The difference is that you were probably not only unaware his save percentage was that low, but you also don't know why that is.

There's no denying that goaltending is extremely difficult to accurately judge based on numbers only. Wins and losses are terrible indicators - just ask Brian Boucher about that - while goals against average is also incredibly influenced by the team in front of you. Save percentage is generally recognized as the best way to judge a netminder, and that's certainly true of the three just listed. But save percentage doesn't account for a team who takes an inordinate amount of penalties (like the Flyers) or teams with poor penalty killing.

For this reason, even-strength save percentage is becoming a preferred method of evaluating goaltenders from across the league. Because teams have a higher shooting percentage on the powerplay than at even strength, goalies will have a lower save percentage on the penalty kill. This is not the fault of the goalie, but a percentage game that depends upon the team he's playing behind.

Right now, Bobrovsky has a 0.922 even-strength save percentage. Last year, Michael Leighton had a 0.916 ESS% for the Flyers, while Jonathan Quick registered a 0.919 percentage and Ryan Miller recorded a 0.928 percentage. It's quite possible that Bobrovsky's ESS% falls below 0.922, but the reason his overall save percentage will improve from 0.904 is due to his save percentage while shorthanded. Currently, Bobrovsky is stopping 79.2% of shots fired while the Flyers are down a man. In contrast, Brian Boucher is at 0.900. That number will likely settle in around 0.870 for both players, meaning Bobrovsky will see his save percentage improve dramatically. As an example, if Bobrovsky stopped 87.5% of shots (or prevented just one more goal) on the PK so far this year, his overall save percentage would be 0.911. So long as Bobrovsky continues playing, his save percentage will likely increase.

Only Going to Fall

1. James van Riemsdyk's plus/minus

The exact opposite of Mike Richards, van Riemsdyk is currently a plus-2 this year, and he's been on the ice for five Flyers goals and only one opposition goal at five-on-five play so far this year. (Here is where you see part of my complaint with plus/minus: empty-net goals and SH goals.) But for the exact same reasons Richards will improve his plus/minus, van Riemsdyk will see his decrease.

Recall that Flyers goaltenders are currently stopping around 91.0% of shots at five-on-five. So far, the goalies have stopped 97.8% of shots while van Riemsdyk is on the ice. Obviously, this can be affected by players based on their defensive skill, but there's no denying that 97.8% is unsustainably high. Last year, Blair Betts led the team with a 93.7% save percentage, suggesting van Riemsdyk will see the puck behind his goalie more often going forward.

2. Sean O`Donnell's plus/minus

Not to keep harping on plus/minus, but the most obvious outlier on the blueline is O`Donnell. Currently, he is tied for best plus/minus on defense with a plus-3. In over 136.5 minutes so far this year, O`Donnell has only been on the ice for one opposition goal, the overtime winner against St. Louis. As a result, Flyers goaltenders have a 100% save percentage at 5-on-5 with him on the ice, which cannot be sustained.

When you factor in that O`Donnell is giving up over 31.3 shots per 60 minutes and starting in his defensive zone 63.6% of the time (most on the team), there will be pucks that find their way behind his goaltender. Based just on how often he's being placed in a difficult spot - his zone start - O`Donnell will see his plus/minus decline. Factoring in how many shots the opposition gets and that none of them have gone in at five-on-five yet, his plus-3 rating is bound to fall.

3. Claude Giroux's goal totals

Just like with Bobrovsky, this one isn't guaranteed to happen, but the numbers suggest they will. Through nine games, Giroux is on pace to score 36 goals this year, or more than double last year's total. While his points-per-game totals are close to his VUKOTA projection (0.78 v. 0.73), he's scoring goals instead of getting assists.

Not only have half of Giroux's four goals come while shorthanded, his shooting percentage is 20%. Nobody in the NHL last year had a 20% shooting percentage.  Giroux may in fact score 64 points this year - especially if he continues getting nearly 19.5 minutes of ice time a game - but he won't register more goals than assists, and he especially won't reach 36 for the year.

Conclusion

Statistics and conclusions drawn through nine games are inherently questionable. Plus/minus, shooting percentage, and save percentage are the three biggest culprits, largely due to the huge variance that occurs when just one goal is scored this early.

As said above, it's possible that Richards ends the season as a minus player, just as it's possible Claude Giroux scores 35 goals. But it's much more likely that the results through these nine games are outliers and everything will revert much closer to normal. So comment Danny Briere and Claude Giroux for their fast starts, but know that they are unable to keep this pace for an entire 82 game season. Just like Mike Richards is unable to stay this defensively suspect and Jeff Carter is unable to stay this snakebitten.