Regression Pt. 2 - HITS, PIM, SB, Shots

[eaglesfan036 4,603]

VHL Regression Part 2

 

In case you haven't done so yet, read my first regression, as I explain how to interpret the Eviews results, otherwise you will probably have no idea what is going on. As a reminder, this is based on final player attributes and player stats for the S63 regular season. For this study, I will examine how each individual player attribute impacts expected hits, penalty minutes, shots blocked, and total shots taken. If you are too lazy to read through this whole thing, go check out my summary at the bottom.

 

Hits

 

 

The most statistically significant attributes that impact hits are checking (1% level), defense (5 % level), discipline (5% level), leadership (5% level), experience (10% level), and skating (10% level). As expected, checking by far is the most important attribute to increase your hit totals, as a one tpe increase in checking increases your total expected hits in a season by about 5. Defense has a strong negative correlation with expected hits, which makes sense considering the STHS definition of how defense impacts your player, " The higher the stat, better he'll play in defense / he'll block more shot / he'll retrieved more free puck." If your player is retrieving more lose pucks or poke checking the puck away from the opponent, that means they possess the puck less often, and there are less opportunities to make a check. Increasing discipline also results in less expected hits, as a more disciplined player won't make as many stupid hits taking himself out of play or to avoid penalty minutes. Leadership has a decently strong positive correlation to increased expected hits, which makes sense to me as a leader would put his body on the line more. Experience has a decently negative correlation to hits, which is not something I can explain as one would think experience would have no impact on expected hits. Lastly, skating is finally useful for something, as it is the second strongest positive correlation for expected hits, which makes sense because your player needs to be fast enough to catch up to his opponents in order to hit them.

 

Penalty Minutes

 

 

The most statistically significant attributes for penalty minutes are checking (1% level), leadership (1% level), discipline (5% level), fighting (5% level), and experience (10% level). There aren't too many surprises here, as checking by far has the strongest influence on penalty minutes, which makes sense as a player who dishes out more hits will accumulate more penalties. Leadership is tied with fighting for the second strongest positive correlation to penalty minutes, and I am not 100% sure why. Perhaps because leadership results in increased hits, therefore increased penalty minutes? It just doesn't seem very leader like to me. Discipline as expected has a strong negative correlation to penalty minutes. Fighting as expected has a positive correlation to penalty minutes. Also, players with more experience see slightly less penalty minutes, perhaps this is because the crafty veterans know how to be sneakier with their penalties, and are wise enough not to go in the box?

 

Shots Blocked

 

 

First off, it should be noted that the adjusted R squared for shots blocked is only 50%, which is much lower than for any other statistic I've looked at, meaning that  only 50% of the variation in the dependent variable, shots blocked, is explained by the variation in the independent variables, which are all the different player attributes. For shots blocked, defense, faceoffs, experience, and leadership are significant at the 1% level, and scoring is significant at the 10% level. Very shockingly, leadership has the strongest positive correlation for shots blocked, not defense! It is interesting to see that leadership actually does something, and it makes sense too as a player with higher leadership would lay his body on the line more. Next up defense obviously would have a strong positive correlation to blocked shots. Faceoffs having a negative correlation to blocked shots makes sense, because your team winning the faceoff means the other team has the puck less and shoots less. Experience has the strongest negative correlation to blocked shots, and my only explanation for that is players with high levels of experience are often very good and on the best teams, and their team often has the most puck possession. Lastly, scoring has a negative correlation to shots blocked, and I don't know why. 

 

Shot Attempts

 

 

For shot attempts, scoring is significant at the 1% level, puck handling and penalty shot are significant at the 5% level, and passing is significant at the 10% level. Scoring obviously results in the most expected shot attempts, no surprises there. Puck handling also makes sense having the second strongest positive correlation in shot attempts. Penalty shots having such a strong negative correlation in shot attempts makes no sense to me, and I'm guessing it's probably due to a very small sample size of people who actually upgraded this stat. Passing having a strong negative correlation to shot attempts also makes sense.

 

Conclusion

 

As a quick recap, in order of importance the following attributes have the greatest impact on these statistics. + means positive correlation, - means negative correlation, ? means I am unsure if totally correct.

 

Hits - Checking (+), Discipline (-), Skating (+), Defense (-), Leadership (+), Experience (-?)

Penalty Minutes - Checking (+), Discipline (-), Fighting (+), Leadership (+), Experience (-?)

Shots Blocked - Leadership (+), Defense (+), Experience (-?), Scoring (-), Faceoffs (-)

Shot Attempts - Scoring (+), Penalty Shot (-?), Puck Handling (+), Passing (-)

 

My biggest takeaway from all of this is that contrary to popular belief, leadership actually does something!  @flyersfan1453 falsely told me that leadership is turned off in the sim, but I have received confirmation from @Will that this is not true. Increasing leadership results in more hits, therefore more penalty minutes. Leadership is also the best way to increase your shots blocked totals, even more important than defense according to the numbers. Outside of that, my next biggest takeaway is if @Gooningitup wants to increase his hit totals, skating should be his next focus after checking. As always, please let me know your thoughts, as it appears to me that the numbers make sense and don't lie. Tune in for my next article, which will see which attributes have the strongest impact on save percentage for goalies. 

Edited January 16, 2019 by eaglesfan036

[eaglesfan036 4,603]

1,029 words - for the weeks ending 1/20 and 1/27

[Enorama 2,038]

5 hours ago, eaglesfan036 said:

Experience has a decently negative correlation to hits, which is not something I can explain as one would think experience would have no impact on expected hits.

 

6 hours ago, eaglesfan036 said:

Also, players with more experience see slightly less penalty minutes, perhaps this is because the crafty veterans know how to be sneakier with their penalties, and are wise enough not to go in the box?

 

Reverse causality?

 

Someone who spends a lot of their points on checking will get more hits and penalty minutes, but might contribute less to their team's success, which leads to them getting fewer experience points across their career?

 

Just throwing it out there.

 

 

6 hours ago, eaglesfan036 said:

Lastly, scoring has a negative correlation to shots blocked, and I don't know why. 

 

Also, I would assume this has something to do with higher scoring individuals having their team do better and therefore the puck spends less time in their end.

[Nykonax 1,564]

Thanks nerd

[Quik 4,113]

Interesting read. Don't think I've ever been in the league at the same time as an engaged @eaglesfan036 before, it's nice to see

[eaglesfan036 4,603]

2 hours ago, Nykonax said:

Thanks nerd

Thanks boom

[Exlaxchronicles 415]

Once again nice read! Your two posts on attribute analysis should be added to the " Create A Player " thread as a must read part of the tutorial.

Edited January 16, 2019 by Exlaxchronicles

[WhiteMocha 33]

I don't suppose you'll be doing this for goalies as well? I'm curious to know how much, if at all, leadership effects us netminders. Just for example.

[eaglesfan036 4,603]

4 minutes ago, WhiteMocha said:

I don't suppose you'll be doing this for goalies as well? I'm curious to know how much, if at all, leadership effects us netminders. Just for example.

Someone didn't read the very last sentence

[WhiteMocha 33]

15 hours ago, eaglesfan036 said:

Someone didn't read the very last sentence

 

Oops. In my defense, it was a lot to read. I wound up skipping the second half of that last paragraph. My bad...

 

Looking forward to the next one!

[Baozi 1,660]

Interesting tho I wonder if we can raise leadership effects?

[Beketov 9,041]

He’re hoping Discipline is a better effect on PIM than it does on Hits and I didn’t just completely neuter myself in an attempt to stay out of the box.