Jump to content

The Gustav Effect: Do I Radiate Good Fortune, or is My Ego Just Really Big?


Recommended Posts

When you see something I've written center-aligned, you know we might be here a while. So, TL;DR: simply being on my team will get you hired into a management role if you want one.

 

Let me explain.

 

The S65 Houston crew was (and is) noted for producing an inordinate number of management (GM and AGM) hires in just one season. You could say this was a fluke. Indeed you could. But what if I told you that the exact same thing proceeded to happen again this season in Mississauga? Though I am not, repeat, not, taking credit for anyone's hire, I'm simply suggesting that statistical significance may be drawn from the data below, and that leaves me no alternative other than to claim that we're seeing an example of the GUSTAV EFFECT.

 

Oh, and one more thing: I have no idea at this point whether or not we have ourselves some good significant data. I'm just writing the article. If it's not significant at all, at least it will be amusing to watch my ego plummet to dry earth. If it is, though, I get to claim I have ultimate development superpowers and forcibly take control of the league.

 

To start, let's take a look at every hire I can remember within the past two seasons, in all my time in the league, in no particular order...

 

VHL GMs:

@Esso2264 (NYA)

*@Advantage (MAL)

 

VHL Assistant GMs:

@Matt_O (MAL, beginning of S66, stepped down after being hired in Vegas, so don't you worry)

@Sonnet (HSK)

@Tate (RIG)

@FacebookFighter (MOS)

@DilIsPickle (NYA)

 

VHLM GMs:

*GUSTAV HIMSELF (MIS, and no I'm not using MSH)

*@Thranduil (HFX)

*@InstantRockstar (SDM)

*@Nykonax (MEX)

@Matt_O (LVA)

@Acydburn (OTT)

@Rayzor_7 (MIN)

 

VHLM AGMs:

*@Radcow (MIS)

@berocka (MIS)

*@uphillmoss (OTT)

@HulkHogan (OTT)

@fonziGG (HFX)

@Acydburn (HFX)

@DMaximus (MIN)

*@MexicanCow123 (SDM)

*@Snussu (MEX)

*@StamkosFan (MEX, albeit for a few minutes only)

@cody73 (YUK)

*@Jubo07 (PHI)

@Jables (HOU)

@FakeJenton (LVA)

*@GlowyGoat (HOU)

 

*S65 hires, per my definition being those hires made before Hounds players, and therefore my S66 sample pool, existed.

 

I'm sure I'm leaving a few off the list, but I stared at every team and tried to remember their hires to the best of my ability.

 

Now, anyway, who here has played with me or for me, at some point along the line? I can say with definite authority that I have direct ties to the following individuals mentioned above, in an undisputed manner...

 

HOUSTON TEAMMATES:

@GlowyGoat

@Jables

@FacebookFighter

 

MISSISSAUGA PLAYER-SLAVES:

@Rayzor_7

@HulkHogan

@cody73

 

MALMO TEAMMATES:

@Matt_O

 

These are, strictly, the names that nobody can dispute. If you want to put these up for debate, here are some more, with reasons why they're somewhat questionable...

 

@Sonnet (was "on my team" in Houston, but was my GM at the time. Originally I called this the Sonnet Effect before all the Hounds hires started popping up).

@GustavMattias (yep, that's me. Can I count myself? Debatable). 

@Radcow or @berocka (I hired these guys myself, which is a minus, but they were also actively applying for AGM jobs around the league, which is a plus).

@Matt_O (his hire in Vegas counts, but does his hire as Malmo AGM count? I'm not 100% sure).

@DMaximus (I never played with him outside of Team USA in WJC, but I did recommend him to Ray).

 

And, if you really, really, wanted to stretch it...

 

@uphillmoss (was GM of Team USA in WJC, a team I played on).

@fonziGG and @Acydburn (my teammates in the SBA, where I'm a very quiet welfare claimer).

 

 

So, what does this all mean, and am I in fact (mathematically) responsible? Let's take a look...

 

Determining statistical significance, for nerds, is the method of, simply put, comparing a set of observed data with a set of expected data (i.e. what should be expected to be observed), and then from there figuring out the probability that the set of observed data was collected by pure chance. Basically, we're going to find out how likely it is that the Gustav Effect is a thing.

 

First off, let's take our best-case scenario. We'll assume that:

-Every single name above, even the huge mental-gymnastics-type of people like Uphill, Fonzi, or Acyd, is a hypothetical benefactor of the Gustav Effect.

-Also, our sample size is 17 teams for those hires made in S65 and 21 teams for those made in S66 (to include VHL teams as well as VHLM), with each team having an equal number of active players. This represents the highest sample size possible, and therefore the lowest expected probability that any given hire was made off of my team. For example, a hire made in S65 is counted as one hire out of 17 possible hires, and one made in S66 is counted as one hire out of 21.

 

For all calculations we do here, we'll be using a chi-squared test (which can be seen in the link above, but in case you missed it and really like reading up about math, here's the link again). I'm counting twenty-nine hires above, with 15 in this best-case scenario being mine. So, what do we have here?

 

Our expected results are these:

-In S65, 11 hires were made. 0.647 of these can be expected to be from teams of which I was a part. In reality, three were made--GlowyGoat, myself, and Radcow.

-In S66, we saw 17 hires. Per the best-case model, we can see that approximately 2.42 of these hires could be expected to be from my teams (if we count Malmo, Mississauga, and Houston. This number would be lower if we counted Malmo and Mississauga, but the best-case scenario is meant to be unrealistic, not downright blasphemous). THIRTEEN of these hires, by the best-case model, can be claimed as my own. 

 

So, our expected numbers are, with apologies for my inability to figure out the formatting:

 

 

  S65 S66
Hired 0.647 2.42
Not Hired 9.625 14.58

 

Chi-squared values are calculated by finding the difference between the observed value and the expected, squaring it, and then dividing that value by the expected. They also add up--for example, if I were to calculate a chi-squared value of 5 in every one of the four boxes we have here, our final value will be 20. But, skipping all the complicated math to spare all of you the pain and suffering of reliving your high school math classes, here we go with the chi-squared values.

 

  S65 S66
Hired 8.56 46.25
Not Hired 0.54 5.05

Final chi-squared value: 60.4

 

For anyone who doesn't understand yet, here, have a table:

index.gif

The r-value in this table, which we can see on the left, is the number of "degrees of freedom" in the test. It can be calculated by subtracting one from the number of rows, as well as subtracting one from the number of columns, in the above table, and multiplying those numbers. Here, r is 1 (for those thinking r = 4, one row and one column are made up of labels only, not actual data. So, we have a 2x2 grid, and therefore r = 1 since 1x1 = 1). 

 

For the obtained chi-square value to be considered "significant" at a certain point in probability, the value must be at or above the value listed in this table, in the row where r = 1. Now, as to what "significant" means here, here's what I'm saying: If the chi-square value is, for example, 4, this falls in between chi-square for 0.05 and chi-square for 0.025 in the above table. So, there is between a 5% chance and a 2.5% chance that the values obtained were obtained purely by chance alone, with no outside factors (or Gustav Effects) having any effect on the results. Our value here? 53.494, far, far, above the required chi-square value for 0.01. So, by the best-case model, there is about a 99.9[insert a bunch more 9s here]% chance that the Gustav Effect is legitimate.

 

But what if we change the criteria, and stack the conditions against the Gustav Effect?

 

Here are the changes that we'll make:

-We will only include the undisputed hypothetical benefactors (my Houston teammates, Malmo teammates, and Mississauga players)

-Nobody I have hired is eligible

-To account (and possibly overcompensate) for the presence of inactives, the sample size will be reduced drastically, to 7 VHL teams, 5 VHLM teams in S65, and 7 VHLM teams in S66. This gives a grand total of a sample size of 12 in S65 and 14 in S66, which we'll take down even further to 10 in S65 and 12 in S66 to account for the fact that some of us are already in management to begin with.

 

So, what do we get?

-One hire was made in S65. Out of the 11 hires in S65, one was from Houston. We can expect about .92 "Gustav hires" in this case, making this not out of the question.

-Six hires were made in S66. Here, we can expect about 3.64 hires from my teams.

 

Behold, our expected values...

  S65 S66
Hired 0.92 3.64
Not Hired 11.08 13.36

...and behold, the resultant chi-square values...

  S65 S66
Hired 0.01 1.53
Not Hired 0.01 0.42

...which gives us a total value of 1.97.

According to the table, if we stack everything against the Gustav Effect, we still see results that show that there's over an 80% chance, at worst, that it's legitimate (you can't see that in the table above, but here is a full table that puts the value for 80% at around 1.6). 

 

So, am I the fountain of youth? You bet. Possibly maybe.

 

Word count: 1595 words, will claim for the next three weeks as I'm on vacation in two. Also tagging @Renomitsu to fulfill my quota.

 

58 minutes ago, cody73 said:

@GustavMattias Hey man, if you keep stroking that ego, you know it's gonna blow up in your face, right? 

 

Are you sure you're talking about his ego...?

Stroking intensifies

 

No, I'm not seriously taking credit for anything. We've just had a particular set of circumstances I can make an interesting article about.

Now that it's Sunday (for me at least), claiming for the week ending 7/21. 

Edited by GustavMattias
  • 2 weeks later...

This may be the first time I've written an extended article and claimed every week of it. I'm ashamed. 

 

Anyway, claiming this bad boy for week 3.

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...