Create Account

S52 Draftee Summary/Analysis
#1
(This post was last modified: 01-17-2020, 12:15 AM by Paq.)

Summary & Analysis of S52 Draftees

Background
I thought it would be interesting to take a look at some of the data I could collect for draftees in the S52 class and break it down. I collected a bunch of data and imported it to a Jupyter Notebook where I did all of my data manipulation, visualizations, and analysis. If you'd like to get a sense of how I did any of this or see the effort I put in for any potential bonus, you can access the dataset and code in this github repository. Also I believe this is eligible for 2x media so please pay me tons of money thank you.

Data Collection
All of the data I collected for this was available on player pages to keep things as simple/efficient as I could. I separated goalies and skaters into different datasets, though this piece will focus only on skaters. I thought it would be most interesting to break things down by position and archetype. Additionally, in my time as a GM in the NSFL I did some analysis of TPE gains over time, so I threw a player creation date column in there as well. So I collected basic player information (minus strengths and weaknesses) and player builds. The columns were:

Code:
Player | Team | Position | Type | Created | TPE | Checking | Fighting | Discipline | Skating | Strength | Endurance | Durability | Puck Handling | Face Offs | Passing | Scoring | Defense | Penalty Shot

Draftee Summary

First thing I looked at were some breakdowns of the class, grouped by their position. Started out by keeping it simple with a breakdown of draftees by position:

[Image: 9sAqJgR.png]

Not much to this. If you consider wingers as a single group, we have 18 centers, 17 defensemen, and 19 wingers. Interestingly, nobody was particularly interested in the left side. I'm not sure if there's any reasoning for this in the sim or if it's just an unconscious right-handed bias. Either way, S52 has no shortage of any individual position. What about the most popular archetypes within those groups, though? I noticed that the two-way defenseman was pretty popular, but wasn't sure if any others stood out.

[Image: fGcFyqO.png]

Overall there was a pretty balanced distribution of archetypes within position. Playmaker was the most popular among centers, two-way the most popular for defensemen, and offensive the most popular among wings. The most popular selection overall was, indeed, two-way as a defenseman. There is a definite skew toward offense and/or balance, with a vast majority of draftees falling into one of offensive, playmaker, sniper, or two-way. With a lot of folks choosing pretty similar structures for their players, I got the idea to look at how unique builds were, but more on that later. Let's move on to TPE, the most important stat. Breaking it down by position:

[Image: Fo2Vqz2.png]

There has been a pretty consistent activity level across all four positions, though left wing again finds itself below the other positions with an average 17.4 TPE lower than the next lowest. So not only are people not really creating as left wings, those who do don't seem to be earning TPE at quite the same rate. Before we move on, let's look at draftee TPE for the class as a whole.

[Image: 0lqaAbP.png]

Among the 54 skaters in this class, 14 have earned between 0 and 10 TPE. The average TPE across the class sits at 209 with a standard deviation of 43. Top end talent is scarce, with only three players having a TPE total more than two standard deviations from the mean. There are plenty of players sitting around the 1 standard deviation mark or above, however, so finding those draftees who will keep earning is going to be key. With that in mind, not all TPE totals are created equal. You could have two players with 240 TPE who created a full month apart (potentially, maybe this exact example doesn't apply). So let's look at an admittedly flawed metric, TPE per Day. Before any table, let me acknowledge the obvious flaw. Because we update in weeks, someone who joins Monday of a week and somebody who joins Friday will have the same ability to earn TPE, but one player has five extra days counted against him.

[Image: ebOa8Du.png]

Interestingly a handful of December creates top this list, with Bale being the most prolific earner in the class by this metric despite gaining just 47 TPE thus far. We also see some of the top end earners show up here, with TPE/day totals that are likely more sustainable as they have maintained them for this much time. Adam Scianna, who tops the class in overall TPE, ranks 5th at just below 5 TPE per day as of today.

Uniqueness

Ok so I've covered some basic overview. Like I said earlier, though, I thought uniqueness of build would be a fun thing to look at. First, let's look at the attributes of the average S52 player.

[Image: Wqsfihi.png]

Not a ton of surprise here, as skating, puck handling, passing, scoring, and defense are the highest ratings on average. Obviously discipline is up there at 62.3, but given that all players start with 62 there, it feels safe to exclude it. Obviously builds are going to be different based on position, however, so I decided to take a look at which players were most different from their counterparts. The methodology wasn't very scientific, but it felt like a quick and easy way to get to what I wanted. Essentially what I did was calculate a z-score for each attribute relative to their position's average, then take the average of the absolute values of those z-scores to find the player with the highest average across all attributes. The tables are super wide, so I'm not going to post images here, but here are the highlights.

Most Unique Centers
Player - Uniqueness Score

Simothy Drunkebird - 1.000
Stirling MacTavish - 0.857
Cyril Vskoč - 0.611

Simothy Drunkebird took this one in a runaway, being one of just a few centers to choose the Power archetype. His most unique attribute is checking, with a z-score of 2.472. Other highlights include Skating (1.474) and Passing (-2.139). Stirling MacTavish had a handful of somewhat unique attributes based on z-score, primarily attributes he ignored relative to others, like Puck Handling (-1.468) and Scoring (-1.934).

Most Unique Wingers
Player - Uniqueness Score

Boris Franciszek - 1.445
Rhys Pritchard - 0.726
Cal Labovitch - 0.692

Boris Franciszek is the most unique player in the class, with his attributes more than a full standard deviation from the mean, on average. This is largely thanks to his investment in Checking (3.537) and Fighting (1.909), as well as his ignoring of Puck Handling (-2.694). Rhys Pritchard's ranking is in large part due to his high TPE total. He's simply been able to invest more in attributes across the board than his winger counterparts. Cal Labovitch is in a similar boat, as the two are built fairly similarly. Pritchard has a bit more invested in Defense than most, while Labovitch has turned his focus to puck handling.


Most Unique Defensemen
Player - Uniqueness Score

Simon Moreau - 0.816
Adam Scianna - 0.788
Sabo Tage - 0.713

Not a ton of creativity among the defensemen. Adam Scianna and Sabo Tage are fairly high because of high TPE totals. Simon Moreau doesn't have any exceptionally unique attributes, but has focused more on Skating (1.274), Puck Handling (1.338), and Passing (1.263) than his defensive competition.

TPE Z-Score and Class Ranking
So as part of the previous calculations I also calculated z-score of TPE relative to position. I wanted to come up with a way to rank players that took positional activity into account. For example, if 15/18 Centers have 155 TPE, the three who don't might be valued more highly than they would based on raw TPE because of scarcity. So the following table is just the class, sorted by TPE z-score. A quick reminder that all wingers were grouped together for this, so despite LW having a lower average as we saw earlier, they and RW will be treated as one for this.

[Image: 3W41NOH.png]

And so we can now definitively, and with no arguments ever, declare Simothy Drunkebird as the best player in this draft class. On a more serious note, we can see (limited) usefulness of measuring this in Cal Labovitch's #2 ranking here. Because wingers haven't quite earned as prolifically as defensemen, Labovitch gets pushed ahead of Adam Scianna, who is tied for 1st in the class in TPE. But there's only so much to this. As we can see Drunkebird sits relatively low in TPE/Day among players with the highest TPE z-score. If you look at the two in concert, along with raw TPE total, it is probably Scianna who would be the class' #1 ranked player. How any of this matters in the draft (hint: it probably doesn't) remains to be seen.

I plan to do something with the goalie data I collected though it won't be quite as fun because there aren't very many, but do keep an eye out for that!

---

1685 Words

[Image: xOtrQ61.gif]
.
Reply
#2

#Goaliehate

#wereplayerstoo
Reply
#3

#Goaliehate

#wereplayerstoo

[Image: capybar22_600.png]
Reply
#4

Ilike

[Image: 60133_s.gif]
[Image: W3cWGCN.png][Image: 3XEVZeB.png]
 Timber Steelhawks norway norway Steelhawks Timber  
Knights  Renegades norway   norway Renegades Knights
Reply
#5

The TPE/day is kinda skewed, since early in the season (and in the pre-season, or even the PO before our first season) the possibilities to earn good amounts of TPE were limited. Basically just AC/Training.

But pretty nice piece overall.

 
Falcons Monarchs Switzerland   Switzerland Monarchs Falcons
[Image: qGhUIfY.png] [Image: dGD5tIx.png]
  


Falcons Monarchs Switzerland   Switzerland Monarchs Falcons
[Image: qGhUIfY.png] [Image: dGD5tIx.png]
  


 [Image: mutedfaith.gif]
Credit for the images goes to @Carpy48, @soulja, @fever95 and @Wasty
Reply
#6
(This post was last modified: 01-14-2020, 09:51 AM by Paq.)

01-14-2020, 02:11 AMMutedfaith Wrote: The TPE/day is kinda skewed, since early in the season (and in the pre-season, or even the PO before our first season) the possibilities to earn good amounts of TPE were limited. Basically just AC/Training.

But pretty nice piece overall.

Ah, I didn't quite realize this. TPE/Day tends to be pretty garbage anyway just because of the way update weeks work, but I wanted a different lens to view the class.

[Image: xOtrQ61.gif]
.
Reply
#7

01-13-2020, 02:46 PMGeekusoid Wrote: #Goaliehate

#wereplayerstoo

I plan to get to the goalies! I have all the data, I just can't quite do as much/the same things with it because the group is relatively small.

[Image: xOtrQ61.gif]
.
Reply
#8

#Drunkebird1OA
Reply
#9

01-14-2020, 10:11 AMEsso2264 Wrote: #Drunkebird1OA
Would be nice for my mock.

 
Falcons Monarchs Switzerland   Switzerland Monarchs Falcons
[Image: qGhUIfY.png] [Image: dGD5tIx.png]
  


Falcons Monarchs Switzerland   Switzerland Monarchs Falcons
[Image: qGhUIfY.png] [Image: dGD5tIx.png]
  


 [Image: mutedfaith.gif]
Credit for the images goes to @Carpy48, @soulja, @fever95 and @Wasty
Reply
#10

01-14-2020, 09:51 AMPaq Wrote:
01-14-2020, 02:11 AMMutedfaith Wrote: The TPE/day is kinda skewed, since early in the season (and in the pre-season, or even the PO before our first season) the possibilities to earn good amounts of TPE were limited. Basically just AC/Training.

But pretty nice piece overall.

Ah, I didn't quite realize this. TPE/Day tends to be pretty garbage anyway just because of the way update weeks work, but I wanted a different lens to view the class.
It's still a quality read.

 
Falcons Monarchs Switzerland   Switzerland Monarchs Falcons
[Image: qGhUIfY.png] [Image: dGD5tIx.png]
  


Falcons Monarchs Switzerland   Switzerland Monarchs Falcons
[Image: qGhUIfY.png] [Image: dGD5tIx.png]
  


 [Image: mutedfaith.gif]
Credit for the images goes to @Carpy48, @soulja, @fever95 and @Wasty
Reply
#11

Oh wow that was really interesting probably the only article I’ve read on here that hasn’t talked about myself in like at LEAST 2 real life years. I loved how analytical and statistically unique it was although I do enjoy the more blog style typing with more opinion and personality in a lot of the articles, and that is how I myself also like to write, it’s nice to see articles that are written differently and come from fresh new takes. 10/10 would buy again but next time in a bigger size the right shoe fit my foot but the left one didn’t

Need a sig :-/
Reply
#12

01-13-2020, 02:43 PMPaq Wrote: Most Unique Centers
Player - Uniqueness Score

Stirling MacTavish - 0.857

[Image: tenor.gif]

@Carpy48, @sulovilen, and @Ragnar are signature gods
[Image: MaverickLegend.gif]

[Image: U66t7Jy.png] [Image: zyyvIFD.png]
Dragonite
Reply




Users browsing this thread:
1 Guest(s)




Navigation

 

Extra Menu

 

About us

The Simulation Hockey League is a free online forums based sim league where you create your own fantasy hockey player. Join today and create your player, become a GM, get drafted, sign contracts, make trades and compete against hundreds of players from around the world.