Markstrom vs Vladar - Deficit Goals Against and playing from behind

[tvp2003]

Apologies in advance for the long read — if you want you can scroll to the bottom to see the actual numbers/analysis.

BACKGROUND:

I couldn’t find the exact tweet but I was intrigued by something that Darren Haynes posted on Twitter recently — about how Vladar had a better save percentage in the first period versus Markstrom. It has also been noted elsewhere that Vladar has been getting more “run support”, but why?

There has also been some recent discussion (specifically by Derek Wills) about how a team doesn’t play better or worse in front of one goalie versus another. I can’t say I necessarily agree with this, but how can you measure it?

With how this team has been constructed and coached, they are much better playing with the lead versus trying to generate offence when trailing. Specifically, we don’t have the offensive firepower to score at will, but if we’re tied (or ahead) and the other team has to take more risks to try and score, we are good at exploiting those openings with our forecheck and creating chances on offence.

So how does this relate to goaltending? In my mind, the trouble with Markstrom this year is that he’s been letting in goals — especially early in the game — that have put the team in a hole and have forced them to play from behind. Meanwhile, Vladar has seemingly been able to keep things closer for longer, allowing Calgary to attack with shot volume and eventually build a lead.

So while their GAA (2.75 vs 2.82), and to a lesser extent save percentage, aren’t substantially different, I wanted to see if there was more to the story because not all goals against are created equal. Case in point, on January 14, Vladar let in five goals against Dallas. However, all but one came after Calgary had built up a comfortable lead in the second period, and they hung on to win 6-5. Two nights later, Markstrom only let in two goals, in a 2-1 loss to Nashville. Both goals against came in the first period, and the Flames could only score once after that. Same thing the following game against Colorado — Markstrom lets in three goals in the first period, and Flames can only muster one goal in response the rest of the game.

So with that, I embarked on a statistical analysis to see who was letting in more:

(1) First period goals to put Calgary behind by 1 or 2 goals, aka “deficit goals against” or DGA;

(2) First period goals in total

(3) DGA for the entire game

I was also curious what percentage of DGA’s each goalie let in as a percentage of their total goals against.

Disclaimer: I came up with these metrics before diving into the data, so you can’t accuse me of fudging the numbers. I was just trying to quantify what I had observed on the ice and the stat sheet.

THE RESULTS!!!

Out of a sample size of 48 games — 32 starts by Markstrom and 16 by Vladar, there is definitely a difference:

Markstrom: 0.781 1st period DGA per start
Vladar: 0.375 1st period DGA per start

Markstrom: 1.125 1st period GA per start
Vladar: 0.750 1st period GA per start

Markstrom: 1.312 DGA total per start*
Vladar: 0.937 DGA total per start

* - Markstrom was pulled in two games so if anything that number is lower than what it might have been

Also: of Markstrom’s 84 GA in total, 42 were DGAs (50%), whereas Vladar had 15 out of 44 (34%).

Just for fun I also calculated:

(1) of Markstrom’s 36 1st period GA, 25 were DGAs (69%), whereas Vladar had 6 out of 12 (50%).

(2) of Markstrom’s 84 total GA, 36 were scored in the first period (43%), whereas Vladar had 12 out of 44 (27%).

CONCLUSION:

All of this seems to support the notion that Markstrom is letting in more goals early in the game, and putting the team in a hole as a result. So it shouldn’t be surprising — for this team in particular — that they’re able to produce more offence while playing in front of Vladar… it’s easier for them to score while they’re ahead (or at least tied). And while there are other variables that might be contributing to this, I think it’s reasonable to conclude that Vladar is giving the 2022/23 Flames a better chance to get the lead and then hold it, versus having to come from behind.

Agree? Disagree? Does measuring DGAs make sense? Either way, kudos to you for getting all the way to the end of this post without hitting the back button.

**********

Notes: I did not include overtime goals as I felt 3-on-3 (or often 4-on-3) play was not as meaningful from the goalie’s perspective. I also excluded shootout goals and obviously empty net goals.

Also, I limited DGA to goals that resulted in being down one or two goals, after that I think score effects impact how the teams play in front of them.

[Northendzone]

The data supports what my eyes are telling me

[Jacks]

Wasn't that long of a read, thanks. I agree that this matches the eye test.

[K1LLswitch]

Great analysis - I had a hard time understanding my thoughts on the goalies and you nailed it with this analysis.

When I look at the goalie stats on moneypuck, the numbers show Markstrom is better than Vladar, which doesn’t pass the eye test, and nothing seems obvious in the advanced stats other than Markstrom is weaker at medium danger shots. It is definitely the timing of their goals against that is the problem.

Well done!

[GranteedEV]

While this is a fair assessment, I have one additional theory:

1) Vladar gets superior opponents
2) Superior opponents spend less time in their own zone, allowing the Flames to generate more rush scoring, and reducing rush scoring against the Flames, since the Flames are also spending less time in the opponent's zone

I don't have the numbers to back it up, but maybe someone could dig into it

[FlamesAddiction]

Originally Posted by tvp2003

CONCLUSION:

All of this seems to support the notion that Markstrom is letting in more goals early in the game, and putting the team in a hole as a result. So it shouldn’t be surprising — for this team in particular — that they’re able to produce more offence while playing in front of Vladar… it’s easier for them to score while they’re ahead (or at least tied). And while there are other variables that might be contributing to this, I think it’s reasonable to conclude that Vladar is giving the 2022/23 Flames a better chance to get the lead and then hold it, versus having to come from behind.

Agree? Disagree? Does measuring DGAs make sense? Either way, kudos to you for getting all the way to the end of this post without hitting the back button.

Yes, it makes total sense. This is why teams that score first typically win 70% of the time on average. When a team is playing with a lead, they can be more patient and lock it down. Conversely, when a team is playing from behind, they have to open up more leaving them prone to mistakes. This is why when the goalie lets in an early goal and the team in front also only scores a couple of goals or fewer, you can say they would have lost anyway because they couldn't score.

The added pressure and urgency as the clock runs down also probably leads to more penalties. I don't have data for that, just a hunch.

Giving up an early lead can, and often does, change the whole dynamic of a game.

[indes]

Great work! Huge respect for putting the time in to find evidence that explains what we're seeing. I would love to see these stats across more teams to get some more context.

Very cool.

[tvp2003]

Originally Posted by K1LLswitch

Great analysis - I had a hard time understanding my thoughts on the goalies and you nailed it with this analysis.

When I look at the goalie stats on moneypuck, the numbers show Markstrom is better than Vladar, which doesn’t pass the eye test, and nothing seems obvious in the advanced stats other than Markstrom is weaker at medium danger shots. It is definitely the timing of their goals against that is the problem.

Well done!

Thanks! I know there are better/advanced stats to measure goalie performance (looking at quality of scoring chances and expected goals against), but I was trying to find something that might help explain what we’re seeing in terms of run support and when goals against are being scored. People often talk about making “timely” saves, and I think with this team those saves need to come early and when the score is tied.

Originally Posted by GranteedEV

While this is a fair assessment, I have one additional theory:

1) Vladar gets superior opponents
2) Superior opponents spend less time in their own zone, allowing the Flames to generate more rush scoring, and reducing rush scoring against the Flames, since the Flames are also spending less time in the opponent's zone

I don't have the numbers to back it up, but maybe someone could dig into it

I’d have to check the schedule and the chart I created, but amongst games against the same opponent, I think Vladar has fared better. Obviously not necessarily apples to apples since other things factor in though (home vs away, rest, lineups, etc).

[GGG]

The question I have is is this real or random chance? With goals be low event a Vladars sample size small is it possible that the flames have been simply luckier vs any difference in his play.

I think to test this you would need to look if DGA is a stat that is consistent year to year for goalies or is it more random?

[DeluxeMoustache]

Quick first impression… Really good post. I like this a lot. Fresh way to try to quantify things that you see.

I know there are stats based on game score, and in this case, you create a hypothesis based on what you are seeing, then figure out how to analyze the data objectively to quantify

Going to let this marinade for a bit. Thanks! Good job

[jayswin]

Excellent analysis and definitely supports the eye test in watching games between the two goalies.

[Freddy]

Thank you for posting this! I don’t hate Markstrom and find this season he has made amazing saves to keep the Flames in games they’re only down 1 or 2 but your analysis explains why they were in that situation to begin with. I’ve been vocal about starting Vladar because he looks far more composed, but as games proceed he does give up questionable goals that are easier to forgive him for if there is a 1 or 2 goal lead. Tough call for the coaching staff this season but for me I’d roll with Vladar while our playoff position is so precarious, until the wins dry up and go back to Markstrom with fingers crossed he can start games better.

[tvp2003]

Originally Posted by FlamesAddiction

Yes, it makes total sense. This is why teams that score first typically win 70% of the time on average. When a team is playing with a lead, they can be more patient and lock it down. Conversely, when a team is playing from behind, they have to open up more leaving them prone to mistakes. This is why when the goalie lets in an early goal and the team in front also only scores a couple of goals or fewer, you can say they would have lost anyway because they couldn't score.

The added pressure and urgency as the clock runs down also probably leads to more penalties. I don't have data for that, just a hunch.

Giving up an early lead can, and often does, change the whole dynamic of a game.

Interestingly, I started looking at this last night and noticed that we were not following this trend at all early in the season. In our first 10 or more games the team that scored first only won like 2-3 times. As the season progressed it definitely shifted back in favor of what you'd expect (there was something like 15 games in a row where the team that scored first won).

Originally Posted by GGG

The question I have is is this real or random chance? With goals be low event a Vladars sample size small is it possible that the flames have been simply luckier vs any difference in his play.

I think to test this you would need to look if DGA is a stat that is consistent year to year for goalies or is it more random?

It's a good question -- I took a stats class in university about 100 years ago so I don't really recall; I thought Vladar's sample size of 16 starts wasn't unreasonable though. We'll see how things progress as the season continues on.

As for a year-to-year analysis, I suspect DGA is influenced by different factors, not just the goalie. For example, I recall last year's team frequently had good starts, and even if they got behind it early wasn't a big deal as our top line could often bail them out later in the game. That is definitely not the case this year. At least with this analysis you can start testing the theory of whether (or why) they seem to play better -- and score more -- when Vladar is in net versus Markstrom.

[Bingo]

I'd love to be able to find xga by period.

Then you'd have either a goalie that's not ready or a team that's not ready as a comparison.

Which would then send you spiraling down the lane with one faction saying the Flames are crappier in front of one goalie because they don't trust him vs one goalie being unlucky because the team isn't as sound in front of them.

[Inferno]

Originally Posted by Bingo

I'd love to be able to find xga by period.

Then you'd have either a goalie that's not ready or a team that's not ready as a comparison.

Which would then send you spiraling down the lane with one faction saying the Flames are crappier in front of one goalie because they don't trust him vs one goalie being unlucky because the team isn't as sound in front of them.

I have noticed lately that on many Vladar starts the other team doesn't get as many shots on net to start the game as they do when Markstrom plays and the chances aren't as dangerous. That definitely helps with not giving up an early goal.

Like you said, you could say it's because they don't trust him or you could say he has bad luck and they decide to not show up when he plays.

[GreenLantern2814]

Trust. The secret is trust.

Vladar is going to let in three goals, almost assuredly. But they’ll all be honest goals.

Markstrom is going to let in three goals, and the bench is going to spend a period and a half wondering how two of them got through, and if there’s more tom####ery to come.

The devil you know vs the devil you don’t.

[Bingo]

Originally Posted by GreenLantern2814

Trust. The secret is trust.

Vladar is going to let in three goals, almost assuredly. But they’ll all be honest goals.

Markstrom is going to let in three goals, and the bench is going to spend a period and a half wondering how two of them got through, and if there’s more tom####ery to come.

The devil you know vs the devil you don’t.

Right ...

But Vladar has a the inferior low danger save percentage, so he's given up more "weak" goals as a straight stat.

So not sure that theory holds.

The first period angle is interesting though. Does Markstrom give up more goals low danger first period goals?

That we don't know.

[Paulie Walnuts]

I think one of the issues with Markstrom and 1st period which is kind of deflating is that the team comes out and controls the play and he sees nothing for 7-8 minutes and the 1st shot on him is usually a breakdown or something and we need that save but it goes in and kind of deflates the team.

I think its the timing of goals.

[GioforPM]

Originally Posted by Bingo

Right ...

But Vladar has a the inferior low danger save percentage, so he's given up more "weak" goals as a straight stat.

So not sure that theory holds.

The first period angle is interesting though. Does Markstrom give up more goals low danger first period goals?

That we don't know.

Originally Posted by Paulie Walnuts

I think one of the issues with Markstrom and 1st period which is kind of deflating is that the team comes out and controls the play and he sees nothing for 7-8 minutes and the 1st shot on him is usually a breakdown or something and we need that save but it goes in and kind of deflates the team.

I think its the timing of goals.

I think the latter jibes with my eye test. It's like they don't give him a chance to field some easy saves.

To me the whole question of "does the team play better in front of X" is kind of an indictment of the team more than the goalies, since, to my eye they are not playing that much differently.

[DazzlinDino]

Originally Posted by GranteedEV

While this is a fair assessment, I have one additional theory:

1) Vladar gets superior opponents
2) Superior opponents spend less time in their own zone, allowing the Flames to generate more rush scoring, and reducing rush scoring against the Flames, since the Flames are also spending less time in the opponent's zone

I don't have the numbers to back it up, but maybe someone could dig into it

Another angle is that the Flames offence struggles to adapt when other teams lock down and play with a lead. Not making excuses for Markstrom, but if our offence can't recover as easily as last season that is not all on Markstrom.