Jump to content

Recommended Posts

Today, I was told that the Gustav Effect does not exist.

 

What? Shock. Horror. Absolute madness. Retching noises. This is unspeakable, and I feel so strongly about silencing the doubters that I'm willing to take another significant chunk of time out of my life to rant about it (though the TPE incentive doesn't exactly hurt). In my previous article on the matter, which dropped about 6 months ago to mixed reviews, I took a look at every hire which has been made around the league since I joined, and ran the numbers to determine statistical significance--in other words, did my presence on a team significantly impact hires into management positions from that team? The article in question determined that under lenient criteria, the Gustav Effect absolutely exists, while under strict criteria, while likely that it exists, there's about a 20% chance that it doesn't and people from my team just happened to be hired. So, I'm writing an update to answer the question once and for all...

 

Does my presence on a team lead to others on that team getting hired as GMs or AGMs?

 

Let's break it down.

 

I actually haven't figured out what's up with the numbers yet at this point; I'll be doing the calculations as I write. So, if this backfires in my face, too bad, I get to claim it anyway.

 

Since I joined the league in S65, here's a list of everyone (as far as I know) who has been hired (officially, sorry Calgary) as a GM or an AGM (and here's something else in parentheses, yes, I know, I do this a lot). Names will be separated and color-coded by season.

 

VHL GM:

Advantage (Malmo)--S66

Enorama (DC)--S67

diamond_ace (Prague)--S67

Esso2264 (New York)--S67

Jubo07 (Helsinki)--S69

Peace (Toronto)--S70

Moi (Davos)--S70

 

VHL AGM:

Matt_O (Malmo)--S66

FacebookFighter (Moscow)--S66

Dil (New York)--S66

HenrikZoiderberg (Prague)--S67

McWolf (DC)--S67

Tate (Riga)--S67

Sonnet (Helsinki)--S67

ColeMrtz (New York)--S68

eaglesfan036 (Moscow)--S69

Dil or Blade (Calgary, whichever one was official)--S69

cody73 (Calgary)--S70

Nykonax (Riga)--S70

 

VHLM GM:

Thranduil (Halifax)--S66

Ya Boi (Mississauga)--S66

InstantRockstar (San Diego)--S66

Nykonax (Mexico)--S66

Frostbeard (Houston)--S66

Rayzor_7 (Minnesota)--S66

Acydburn (Ottawa)--S67

Matt_O (Las Vegas)--S67

fonziGG (Halifax)--S68

DMaximus (Philadelphia)--S68

motzaburger (Mexico)--S69

Poptart (Las Vegas)--S69

Dil (Halifax)--S70

Berocka (Halifax Mississauga)--S70

Elmebeck (Saskatoon)--S70

MexicanCow123 (Yukon)--S70

 

VHLM AGM:

GlowyGoat (Houston)--S65

BOOM (Philadelphia)--S65

Radcow (Mississauga)--S66

uphillmoss (Ottawa)--S66

MexicanCow123 (San Diego)--S66

StamkosFan (Mexico)--S66

Snussu (Mexico)--S66

Jables (Houston)--S66

Acydburn (Halifax)--S66

McWolf (Halifax)--S66

FakeJenton (Las Vegas)--S67

DMaximus (Minnesota)--S67

Berocka (Mississauga)--S67

motzaburger (Mexico)--S66

fonziGG (Halifax)--S67

cody73 (Yukon)--S67

HulkHogan (Ottawa)--S67

Poptart (Las Vegas)--S68

Mr_Hatter (Minnesota)--S68

Elmebeck (Saskatoon)--S68

Cxsquared (Mexico)--S68

JeffD (Halifax)--S68

Krice13 (Yukon)--S69

Alex Bridge (San Diego)--S69

DoktorFunk (Mississauga)--S70

fever95 (San Diego)--S70

 

 

So NOW IT'S TIME TO RUN THE NUMBERS. To find out what affects what, we need to look at three factors:

 

-The number of hires made, in total

-The number of hires made, from my teams

-The number of teams in the league at the time the hire was made

 

So, note the differences! For hires made prior to the formation of expansion teams (GM and AGM hires of said expansion teams included--in other words, any hire made before that team had a roster to pick from), those hires have been represented in the color of the previous season, for my convenience and for your confusion. This is because our third factor must be taken into account--in S67, for example, the league was brought up to its current size with the addition of Prague and DC--but people hired to those teams, as well as people hired to any team prior to said expansion, are mathematically equivalent to a hire from S66, as the expansion teams did not have a roster to pick from (and, therefore, could not have seen any hires being made from them) at that point in time.

 

Now, a couple rules must be put into place to ensure that I'm not sneaking people in for my own benefit...

-The Gustav Effect can only apply to players with no management experience prior to appearing on my team. This, for example, rules out Nyko, who was hired as AGM in Riga while playing for the Hounds, but who has previously been both GM and AGM for other teams.

-Myself being hired, as well as people being hired by me, do not count as someone who has been influenced by the Gustav Effect, but they do not count as someone who hasn't. In other words, these situations will be treated as if a hire never happened in the first place. I believe this is fair as, though I was the one who directly controlled the "hired" status of these people, each one had been actively applying for positions around the league prior to being picked up by me (this eliminates Radcow in S66, Berocka in S67, DoktorFunk in S70, and myself in both S66 and S70). 

 

For this we'll be using a chi-squared test, which you can read all about here (Look it up and do your own reading on the matter if that doesn't explain it well enough). A chi-squared test determines the probability that a certain set of results is due to chance, without any outside influences (read: Gustav Effects). The formula can be found below...

 

chi-squared-equation.png?mtime=201510232

 

...and, basically, the higher the number is, the greater the chance that we've got us a Gustav Effect.

 

The test also relies upon "degrees of freedom," or, in other words, the number of choices between possible outcomes. In this case, we've only got two outcomes--hired and not hired, so that means we've only got one possible difference in outcomes, or one degree of freedom. Here's a table:

chi-table_med.jpeg

 

We're only going to be looking at the top row here, for one degree of freedom. But, as you can see on the bottom, "significant" values are those with a 5% chance of being randomly stumbled upon--or a 95% chance that these values were influenced by other factors. So, if the numbers we end up with come out to 3.84 or higher, in total, then we've got us a Gustav Effect.

 

Let's start by looking at who I can claim as a Gustav Effect hire...

 

HOUSTON TEAMMATES:

GlowyGoat (Houston)--S65

Jables (Houston)--S66

FacebookFighter (Moscow)--S66

 

MALMO TEAMMATES:

Matt_O (Malmo)--S66

Matt_O (Las Vegas)--S67

fonziGG (Halifax)--S68

 

MISSISSAUGA UNDERLINGS PLAYERS:

Rayzor_7 (Minnesota)--S66

cody73 (Yukon)--S67

HulkHogan (Ottawa)--S67

ColeMrtz (New York)--S68

 

Cxsquared (Mexico)--S68

JeffD (Halifax)--S68

Krice13 (Yukon)--S69

cody73 (Calgary)--S70

Berocka (Halifax Mississauga)--S70

 

DEDUCTIONS (People who are either me or hired by me)

Radcow (Mississauga)--S66

Ya Boi (Mississauga)--S66

Berocka (Mississauga)--S67

Moi (Davos)--S70

DoktorFunk (Mississauga)--S70

 

So, now we need to break down the number of teams in the league, and when they were in the league.

-In S65 there were 8 VHLM teams and 9 VHL teams, for a total of 17.

-In S66 there were 11 VHLM teams and 10 VHL teams, for a total of 21.

-From S67 to S70, there have been 11 VHLM teams and 12 VHL teams, for a total of 23. 

 

We'll reduce the number to the following, to account for a) teams largely made up of inactives, and b) people already in management.

-13 total teams in S65

-16 total teams in S66

-18 total teams from S67-present

 

So, under the "13-team" umbrella of players hired in S65 and players hired in S66 prior to the formation of expansion team rosters, we've got:

GlowyGoat (Houston)--S65

 

In the "16-team" space of hires made in S66 and before the formation of expansion team rosters in S67, we've got:

Jables (Houston)--S66

FacebookFighter (Moscow)--S66

Matt_O (Malmo)--S66

Rayzor_7 (Minnesota)--S66

 

And, in the "18-team" space from S67 to today, we've got:

Matt_O (Las Vegas)--S67

cody73 (Yukon)--S67

HulkHogan (Ottawa)--S67

fonziGG (Halifax)--S68

 

ColeMrtz (New York)--S68

Cxsquared (Mexico)--S68

JeffD (Halifax)--S68

Krice13 (Yukon)--S69

cody73 (Calgary)--S70

Berocka (Halifax Mississauga)--S70

 

Counting the total number of hires above, we can see that we've had:

-11 hires made in the "13 team" group (13 hires total, but minus myself and Radcow)

-15 hires made in the "16 team" group

-30 hires made in the "18 team" group (33 total, minus myself, Berocka, and DoktorFunk)

 

And out of the group that I'm (jokingly, mind you) taking credit for, we have:

-1 hire in the "13 team" group

-4 hires in the "16 team" group

-10 hires in the "18 team" group

 

I'm considering myself to be a member of Houston only in S65, and a member of both Mississauga and Malmo in every other season. So, that's one team in S65, and two teams everywhere else. To stack things against myself, even, I'll consider both of my teams full teams, with no deductions due to the presence of inactives or other members of the team who were in management.

 

So, we can expect:

0.846 hires to have been made from my team in the "13 team" group

1.875 hires to have been made from my teams in the "16 team" group

3.333 hires to have been made from my teams in the "18 team" group

 

So, now let's plug everything into the above equation. With apologies for the terrible formatting as I have no idea how to center this, here's the table of expected values for S65:

 

  S65 S66 S67-Present
Hired 0.846 1.875 3.333
Not Hired 10.154 13.125 26.667

 

And, the chi-squared values...

 

  S65 S66 S67-Present
Hired 0.028 2.408 13.336
Not Hired 0.002 0.344 1.667

 

These values are cumulative, meaning that the sum of each value provides the total value. Adding these up, we get a final chi-squared value of 17.785.

 

According to the table above, reaching a point where it is 99% probable that the Gustav Effect exists would require a value of 6.63--far below what we've reached here. Basically we've got a virtual proof of the Gustav Effect's existence--or, at least, a 99.999 (keep going with some more 9's) percent chance of it. In fact, it would take 10 degrees of freedom to make these results insignificant.

 

So, the Gustav Effect exists. GAZE UPON MY MIGHTY AURA, YE PEASANTS, AND QUAKE.

 

 

1,639 words, good for three weeks. Obligatory @Renomitsu tag here.

Link to comment
https://vhlforum.com/topic/76263-the-gustav-effect-2-electric-boogaloo/
Share on other sites

seems like a lot of numbers that i dont understand. Stamkosfan was also my agm for about 2 hours, so you can leave stamkos out of that tbh. I say because of that these numbers are bullshit and need to be recalculated

Just now, Nykonax said:

seems like a lot of numbers that i dont understand. Stamkosfan was also my agm for about 2 hours, so you can leave stamkos out of that tbh. I say because of that these numbers are bullshit and need to be recalculated

Fun fact: if I left stamkosfan out of this, it would work even more in my favor.

  • 3 months later...

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...