Login

***Revontulete*** · 03-29-2024, 03:09 AM

Introduction
Hello Everyone! I only joined on Sunday, know very few people, and have been trying to catch up on much of what all I need to do to participate. I wished for some TPE so chose to tackle the mock draft for SMJHL a bit more statistically. I can recognize that many of the posts there may be copy pasted with slight changes, or by many players that are not as active. Yet, it is with my struggle to cipher information in the #to-kill-a-mockdraft channel of the S77 Draft Class Discord Server and it's 1,000 messages that I've turned to the official guesses of players that have posted in the draft thread so far. Due to this limitation of data, everything discussed within this post shall be limited to public information only.

The Approach
I broke this down into several different steps. Allow me to outline them here for transparency sake:

Step 1: Defining the problem statement or hypothesis of the data.
Step 2: Collecting the data for the problem statement.
Step 2.1: We'll be gathering what is called first-party data as all of it shall come directly from the posts in the main draft.
Step 3: Cleaning the data. In other words, removing unwanted data points.
Step 4: Analyzing the data. It is here where the patterns of the data that emerge become important.
Step 4.1: The analysis will be descriptive in nature, as it will be used to create my own mock draft.
Step 5: Sharing my results. This is what I am doing in this post.

Disclaimer
Due to my unfamiliarity with the SHL, SMJHL, the communities inside, and my peers within the S77 Draft Class, many inferences are inherently flawed. Much of the publicly available data comes from various levels of educated assumptions. It is due to this limitation, last minute trades, and more that many of these mock drafts are nonsense at best. While TPE is an encouraging payment for new players, it's like making a March Madness bracket.

This information is based off of slightly changing data as well, meaning that these data points could be entirely different by the draft tomorrow.

I'm also recovering after surgery from yesterday and putting this together in a day, so there's no way half this data is fully coherent. Either way, enjoy reading the ramblings of... this ♥.

Step One: Problem Statement
Which draftees will end up in which position in the draft on the 29th? Much data is obscured behind the scenes, but much can be obtained by the lists created and specifically who creates those lists. This is the basis for how I shall create my own mock draft.

Step Two: Data Collection
To ensure the strongest data collection possible, I split up prioritization of data into a few different sections. The first would be glancing into where the individual players put themselves on their mock draft. For example, the person with the player McLovin put their individual ranking at #4 of their mock draft list. The train of thought with this choice is to recognize that many draftees have a deeper understanding on which GMs they meshed with best than outside actors (other draftees) trying to choose with a lack of information. The image below is the collection of that data organized by the rank they've placed themselves.

https://i.imgur.com/N1W7qPj.png

The second piece of data prioritization will be how often an individual was put into a position. For example, the player Smitty Werbenjaegermanjensen has the plurality of the 17th section of the thread with 8 posts out of 17 posts. This lead compared to other players chosen for that position Smitty Werbenjaegermanjensen as the leading pick of #17 overall. This will act as supplementation for ranks that are not chosen by the previous data set, and in rare cases overrides the position that the players choose for themselves. The image below is the collection of data organized by the number of times a player has shown up per position throughout the 22 spots.

https://i.imgur.com/8gPqa71.png

The third piece of data prioritization is whether or not a player is a rookie or a recreate. I've heard dozens of times since joining that an active rookie is more valuable than a recreate. As this analysis won't rely on data within the S77 Draft Class discord, it will assume that a rookie is active. This is because interactive games and forums often encourage activity more than lurking. This data will be displayed with three different colours. Blue for rookies, yellow for recreates, and white for players that are undefined. (I am colourblind, so the colour determination may be inaccurate. Just categorize the displayed colours with what's closest. ie: blue and purple are closer visually than blue and green... I think). The image below is the same collection of data before, except with the individual players highlighted to represent what's mentioned earlier in the paragraph.

https://i.imgur.com/L0yaxl1.png

Step Three: Cleaning the Data
This is where the previous data is pruned. I already organized them in numerical order for you to easily digest the data, but there is a bit more that must be done. For example, three players listed themselves as the 1st overall rank. Swedish Chef will become the first name on the list because they put themselves 1st overall AND had 14 out of 18 votes from the thread in total. Another problem is when you view ranks. Because this is a glorified guessing game, There were individuals that appeared as the plurality vote for ranks several times. These are Elvar Gil-Galad, Theo Kane, Trevor Lopez, and Olafur Atlason. Their weight compared to the runner up, whether or not they are a rookie or recreate, and where they individually voted will determine where they will land on the overall list.

Step Four & Five: Analyzing the Data and Results
Analyzing the data from here will lead to the supposed definitive list based on the weighted values listed before. This list is
1. Swedish Chef
2. Skyler Stevens
3. LiMu Emu
4. McLovin
5. Emeric Lavoie
6. Simon Science
7. AT-AT Wollker
8. Elvar Gil-Galad
9. Olafur Atlason
10. Theo Kane
11. Trevor Lopez
12. Sam Volta
13. Doug Weight
14. Gwendolynn Telenn
15. Roquefort Cotswald
16. Crazy Tomato II
17. Smitty WJMJ
18. Maj O'Nayse
19. David Vent
20. Kobe frobe
21. Max Kielinen
22. Hockey Player

The logic per rank goes by:
1. Explained beforehand
2. Skyler Stevens voted themselves as 1st overall, and had a plurality of the votes
3. LiMu Emu had a majority of the votes with no individual ranking
4. McLovin voted themselves 4th overall and appeared several times throughout the player votes, unlike Marian Hanak.
5. Emeric Lavoie gained a majority of the votes, and 5x more than Emeric Gagner for the section. While individual ranking has a higher weight, a five time vote shows a popularity that increases the chances of a higher rank.
6. Simon Science gets the ranking despite Theo Kane's lead due to being a rookie and individually voting. This can be indicative of the player knowing they'll be picked by the 6th draft pick as a new player.
7. AT-AT Wollker showed up several times throughout the data set, and with an individual ranking of #7 as well as second place in the overall player votes, they are a comfortable pick.
8. Elvar Gil-Galad lands in 8th as their individual ranking was here and they had the plurality of overall votes. This is despite Gwendolyn Telenn's status as a rookie.
9. Olafur Atlason lands this rank due to individually ranking themselves at 9th and as the lead vote overall. Despite Sam Volta's status as a rookie, Olafur Atlason's continuous appearance throughout the votes tips the scales compared to Sam Volta
10. Theo Kane lands here due to being a popular favourite of the player votes, showing up 3 different times , while having no other player to fight for this spot.
11. Trevor Lopez voted for this rank, and due to voting popularity landing them plurality in two other seconds and being a rookie, this will be a rare time where a ranking occurs despite the lack of votes in the individual rank for the player.
12. Sam Volta lands here as they ranked higher in the player votes to the second place and is a rookie
13. Doug Weight lands in this spot as Theo Kane was placed higher and Doug Weight is the highest ranking rookie of the 13th place
14. With Trevor Lopez and McLovin ranking higher, Gwendolynn Telenn ends up as the last choice for the 14th rank.
15. Roquefort Cotswald was the individual ranking and highest picked by the mock drafts for #15
16. Crazy Tomato II lands here as they were a mock draft favourite beforehand and are the lead for this section.
17. Smitty WJMJ lands here as they are a generally popular pick, and while they didn't land in 15th due to Roquefort Cotswald, can land comfortably in the lead as a recruit ahead of May O'Nayse.
18. May O'Nayse is a popular vote, being barely beaten by the rookie Smitty WJMJ in #17. With the runner up for $17 being two votes down, May O'Nayse fits best here.
19. David Vent is the only qualified for this rank
20. Nor Ge lands here as their individual ranking was here and they were right behind May O'Nayse in the 18th rank
21. As the first qualified player, and a rookie, Max Kielinen comfortably lands here
22. Hockey Player becomes the position here due to previous votes and having no individual ranking in this position.

***Seany148*** · 03-29-2024, 03:19 AM

:D

***Inf1d3l*** · 03-29-2024, 06:47 AM

Really good job with this, was a great read!

Tripl3tsGamer · 03-29-2024, 09:46 AM

Really well explained. I like it!

ec06aaj · 03-29-2024, 09:48 AM

oh it's data analysis, oh it's beautiful

NONAME · 03-29-2024, 03:46 PM

This was a good read! Pay the man.

Navigation

Extra Menu

About us