Elite Dawg Say...

[Tdawg]

So the first annual EDS is in the books, and we have all the information you need to see how this season will finish. We had good participation with 32 people participating in this event--which gives us statistical credibility for the data. Bear in mind, I'm not hear to offer personal opinions--other than making fun of stupid projections. I'm simply here to present the data in a meaningful way.

At the heart of this projection is a simple mathematical exercise of averaging the Percent Chance that ED says we have of winning each game, summing those up, and coming to a projected win total. Below, you'll see the percent chance that ED says we have of winning each game:
Southern Miss - 93.8%
UA Birmingham - 93.6%
@ S. Alabama - 89.3%
@ LSU - 36.7%
Texas A&M - 64.2%
Auburn - 53.5%
@ Kentucky - 73.8%
Arkansas - 73.7%
UT Martin - 97.0%
@ Alabama - 19.4%
Vanderbilt - 76.3%
@ Ole Miss - 52.0%

So, as you can see, ED feels we have a greater than 50% chance of winning 10 of our 12 games. Sounds wooly, right? Well, win you sum the percent chance of winning, Elite Dawgs official season projection is 8.2 wins. So, while the group, as a whole, feels really good about our chances in most games, we do understand that shit happens and the 10 wins will likely not come to fruition. What is interesting is that I would think it unusual for the number of "projected" wins to be different from the number of "pure" wins by more than 1 game. What does this mean? It kind of means that the season is set up for disappointment--not necessarily because 8, or even 9 wins would be a disappointment, but simply because we fee like every game other than LSU and BAMA is a "should win" type of game. What is also interesting, when you put this information in the context of the offseason message board mantra that Dan "needs to win a big one", is that our fans seemed to have moved the bar on him. In only seeing 2 games as "should lose" games, that really limits the opportunities for upset type wins--at least in the minds of the fanbase.

When you begin digging further into the projections, we find that there are 7 games where 100% of the participants give us a >50% chance of winning. Those are the 4 NC games plus Kentucky, Arkansas and Vandy. Furthermore, 94% of the participants gave us a >50% chance of beating Texas A&M. That's 8 games right there where, if we lose one of them, the participants in this forum will see the result as an extreme disappointment. Only 2 games had somewhat significant split among fans regarding whether we "should win" or not. Auburn and Ole Miss. Those games represent the swing games in the season. ED feels outstanding about having 8 wins as the floor to the season and a win in one or both of the games against Auburn and Ole Miss will facilitate the opportunity for an outstanding season. Lose both of those, with "should lose" and "will lose" games vs LSU and Alabama, and all of a sudden there is no room for error throughout the rest of the season.

The Standard Deviations of the data were relatively small across all games except 3. What this means is that there is relative concensus among ED participants that the projections are sound, with the exception of the games vs LSU, Ole Miss and Auburn--in that order. In our mind, the LSU game could go anywhere from a blowout loss to a close win. The Ole Miss and Auburn games could go, literally, in any direction for either team. So, as has been the theme, the only real opportunity for a signature win, according to ED, is vs LSU and, on the flip side, the games that will determine the fate of the season are vs Auburn and Ole Miss. Ole Miss, probably to a lesser degree since so much of the season will have played out to that point and the season will, for the most part, be defined.

When looking at the summed Standard Deviation for all games, Elite Dawgs can say, with 95.4% certainty, that we will win no fewer than 6 and no more than 10 games this season. This is somewhat sobering when put into the context of there being 10 games that the participants feel we "should win". We can also confirm that the odds of us winning 6 games is very low, and the most likely win total range for the season is 7-10 games. When you interpret the data in the context of this message board, you can safely say that 7 wins IS a disappoinment, and 10 wins IS a HUGE accomplishment. The season seems to be destined, however, for either 8 or 9 wins.

Given that there is less than a 25% chance of us losing to both Auburn and Ole Miss, when looking at data provided, the Official Elite Dawgs prediction is 9-3

Elite Dawg Sheep Award

starkvegasdawg - starkvegasdawg has projected 9.1 wins, has us with a greater than 50% chance of winning 11 of our 12 games and only gives us less than a 70% chance of winning in 2 games--LSU (54%) and Alabama (14%). Somebody needs to buy him a gift card to the depression specialist as there's a very good chance starkvegasdawg will be down in the dumps by the time this season ends.

Bear in Dawg's Clothing

Dawg496 - Dawg496 has projected only 7.0 wins and gives us a greater than 50% chance of winning only 7 of our 12 games. With games versus Auburn, Ole Miss and Texas A&M tagged as "should lose" and giving us a very small chance versus LSU and Alabama, Dawg 496 is trying his best to set this season up for achievement

[PassInterference]

Elitedawgs remembers. Outstanding work.

[Political Hack]

outstanding.

[BeastMan]

That is awesome

[Coach34]

And posts like this remind people why we are the best State board

[Alldawg]

Are you trying to say the Gene lacks in math skills>>>>

[DudyDawg]

Most surprising to me is that people are more confident in beating Auburn than TSUN

[starkvegasdawg]

It's ok. If things are worse than I anticipate I have my coping mechanisms:

[CadaverDawg]

What an awesome breakdown. Outstanding work. Makes my prediction look par for the ED course, I guess.

Is there any way we could get you to do some sort of weekly data breakdown for every game based on the board's predictions? I don't know how that would work, but it would be awesome. Well done, Tdawg!

[BulldogBear]

That was very interesting and fun. Is it too late to one on the Bears before Thursday? I'd love to show it to some Bearsharks at work what ED projects for them..

[Political Hack]

That was very interesting and fun. Is it too late to one on the Bears before Thursday? I'd love to show it to some Bearsharks at work what ED projects for them..

that's a great idea if Tdog will do it and has time.

[CadaverDawg]

That was very interesting and fun. Is it too late to one on the Bears before Thursday? I'd love to show it to some Bearsharks at work what ED projects for them..

Great idea

[sleepy dawg]

AMAZING!

Next time this guy asks us to post something and follow his rules, just do what he says without question. Whatever I can do to help this guy make posts like this, I will do.

Thanks Tdawg.

[Tdawg]

That was very interesting and fun. Is it too late to one on the Bears before Thursday? I'd love to show it to some Bearsharks at work what ED projects for them..

Not really gonna have time as I'll be out of town for work most of the week. Y'all could put together a pretty simple one by posing the same question that I did, plugging the answers into an excel spreadsheet, averaging the answers, and summing the averages. It's not terribly time consuming but requires an hour or so in front of my work desktop and then a little time to type out post. You could get the projected win total and "pure" win total pretty easily though.

[Dawg496]

Haha, sorry to be a pessimist. I guess I should be pleasantly surprised if the ED consensus comes true.

[Tdawg]

So the first annual EDS is in the books, and we have all the information you need to see how this season will finish. We had good participation with 32 people participating in this event--which gives us statistical credibility for the data. Bear in mind, I'm not hear to offer personal opinions--other than making fun of stupid projections. I'm simply here to present the data in a meaningful way.

At the heart of this projection is a simple mathematical exercise of averaging the Percent Chance that ED says we have of winning each game, summing those up, and coming to a projected win total. Below, you'll see the percent chance that ED says we have of winning each game:
Southern Miss - 93.8%
UA Birmingham - 93.6%
@ S. Alabama - 89.3%
@ LSU - 36.7%
Texas A&M - 64.2%
Auburn - 53.5%
@ Kentucky - 73.8%
Arkansas - 73.7%
UT Martin - 97.0%
@ Alabama - 19.4%
Vanderbilt - 76.3%
@ Ole Miss - 52.0%

One thing I forgot to do--there is sufficient data on the internet to allow a conversion from Percent Chance of Winning to point spread. If you apply the percents given by ED members to the point spread equivalent, then you would get the following preseason spreads:

Southern Miss: -22.5
UA Birmingham:-22.5
@ S. Alabama: -17.5
@ LSU: +4.5
Texas A&M: -5
Auburn: -1.5
@ Kentucky: -8.5
Arkansas: -8.5
UT Martin: -29
@ Alabama: +12
Vanderbilt: -10
@ Ole Miss: pk

[CadaverDawg]

One thing I forgot to do--there is sufficient data on the internet to allow a conversion from Percent Chance of Winning to point spread. If you apply the percents given by ED members to the point spread equivalent, then you would get the following preseason spreads:

Southern Miss: -22.5
UA Birmingham:-22.5
@ S. Alabama: -17.5
@ LSU: +4.5
Texas A&M: -5
Auburn: -1.5
@ Kentucky: -8.5
Arkansas: -8.5
UT Martin: -29
@ Alabama: +12
Vanderbilt: -10
@ Ole Miss: pk

Wow, that seems really accurate

[LiterallyPolice]

Great Job Tdawg! I love stuff like this.

I used your percentages to create a probability schedule for number of wins. (For those interested, I did a Monte Carlo simulation of 50,000 trials)

Wins Probability
0 0.0%
1 0.0%
2 0.0%
3 0.0%
4 0.4%
5 2.1%
6 8.0%
7 18.4%
8 27.7%
9 25.5%
10 13.7%
11 3.7%
12 0.4%


Probabilty 8 or 9 Wins: 53.2%
Probability of 8 or LESS Wins: 56.6%
Probability of 9 or MORE Wins: 43.4%

[Tdawg]

Wins Probability
0 0.0%
1 0.0%
2 0.0%
3 0.0%
4 0.4%
5 2.1%
6 8.0%
7 18.4%
8 27.7%
9 25.5%
10 13.7%
11 3.7%
12 0.4%


Probabilty 8 or 9 Wins: 53.2%
Probability of 8 or LESS Wins: 56.6%
Probability of 9 or MORE Wins: 43.4%

Here's another guy that knows how to use excel to his advantage. I nominate LiterallyPolice to conduct the Bear Study.

[Jack Lambert]

Most surprising to me is that people are more confident in beating Auburn than TSUN

I think it has more to do with the location.