FB: Old Dominion Athletic Conference

[jknezek]

Yeah. I think Quakers' fans can rest easy on the decision from last year. It is going to be an interesting year, again, in the ODAC. Not sure there is another league that has 4 or more teams with a legitimate expectation to compete for the conference title this year.

On the other hand, it doesn't bode well for our champion, whoever it is, getting a good seed come tourney time.

[abnrgr]

Frostburg had a fatality and it sounds like a new staff might be in order. Drop off your 18 year old and a week later they tell you to extract his body. My God the parents must be numb!!

[BCQBFAN]

Guess who's back?  Back Again!

Hey, skoaltrain!  We need some BC faithful to stir things up.  lol

We've been quiet lately.  It's funny how that works.  Looking forward to a great season.  I know Bridgewater reported 161 to camp.  Quite the numbers.  I need some Bridgewater help.  I can't do it myself.  I've been lurking for a number of months.  I'm back Baby!!!!

I'll try to help some.  Of the 161, 9 quit and at least 20 are day-to-day with various injuries.

[Pat Coleman]

Frostburg had a fatality and it sounds like a new staff might be in order. Drop off your 18 year old and a week later they tell you to extract his body. My God the parents must be numb!!

It happened two years ago. It's jus tthe lawsuit that's new.

http://www.d3football.com/columns/around-the-region/east/2011/derek-sheelys-legacy

[abnrgr]

PC

well once again I am on top of the situation. Next time I do that it should be more than a simple caneing

[jknezek]

I decided to have some fun with the Kickoff rankings. Why? Because I could and because they hammered my Generals (although perhaps not without cause given the holes W&L needs to fill). And so, not because I think I could do better, but simply because this is a message board and such things are made for people to come up with statistically insignificant data that proves a point they want to make, I present the following findings.

Using the 7 teams that have been in the ODAC since Kickoff started doing comprehensive D3 rankings, the most undervalued team given the final season standings relative to the preseason Kickoff is: W&L. The most over-rated team is: E&H.

The complete numbers are here:
W&L   0.625
Catholic   0.5
Guilford   0.375
RMC   0.25
HSC   -0.125
Bwater   -0.5
E&H   -1

Basically if a team has a positive figure, it is under-valued. If it has a negative value, it is over-valued. How did I figure this out? I lined up each of the ODAC teams by the Kickoff ranking and gave 1pt for the highest ranking each year and 7 points for the lowest ranking. I then did the same for the actual finishes. Subtracting the actual from the projected gave me a positive or negative number which was indicative of the total amount of over/under valuing. That number was then averaged over 8 seasons.

So, what does this mean? Pretty much nothing. E&H and Bridgewater tend to be given the benefit of the doubt, W&L and Catholic get a slightly jaundiced eye. HSC is the easiest team to predict, followed by RMC.

What other fun facts can be gleaned? 2011 was Kickoff's best year. They correctly predicted every team in the conference except flipping W&L and HSC, each with a +/-1. Kickoff's worst year appears to be 2008, however 2008 is a tough year to really set the order since 4 teams went 4-2. RMC went to the playoffs and is recorded as the conference champion, but it's hard to fault Kickoff for such a screwy year. Kickoff also got every team wrong in 2006 and 2010, and those years are much more cut and dry. Kickoff has never missed by more than +/-5 with any team, but it was only a 7 team conference. The biggest misses were E&H in 2007, picked to finish 2 and finished 7th, and W&L in 2010, picked to finish 6th and ran the conference.

As far as trends, E&H was hardest to predict in 2005-2008, where they were a +/-3 or greater every year. Since then they have been +/-2 or lower. With 2009 and 2011 being the best years, 2008, 2010, and 2012 being some of the weaker years, it is possible that the rankings are not paying enough attention to the home/away schedule when doing the rankings. That trend, however, bodes well for this year's Kickoff rankings.

Sadly, that does not bode well for W&L, unless, of course, they are once again the most under-valued.

*** First, all records and figures come from either Kickoff or D3football.com site with some collaboration from the ODAC official site. 2008 D3football shows Catholic at the top of the table, but RMC is recorded by the ODAC as the official Champion. Data has the potential to be skewed by any number of ties in the middle of the pack each year by record. I was not going to try and sort them out and, with the exception above, went by the D3football.com table.

*** Second, this isn't a real critique of a group of guys who do a great job every year coming up with 230 or so rankings across D3. Expecting accuracy, or spitting out anything someone might consider a legitimate complaint when they are not accurate, is ridiculous. I did this simply to have some fun on a slow day.

*** One last note. I'm not posting the raw data here because the actual rankings are proprietary to Kickoff. If you want to see it, go spend $11 and have your own fun!

[Pat Coleman]

The most over-rated team is: E&H.

The complete numbers are here:
W&L   0.625
Catholic   0.5
Guilford   0.375
RMC   0.25
HSC   -0.125
Bwater   -0.5
E&H   -1

Interesting. Keith and I often have discussions that center around, "are we overvaluing our alma maters" -- especially in years where they are further down the table. This suggests we're not doing that, at least.

This means even as we overrate E&H or underrate W&L, it's been by an average of one spot per year? I'm OK with that.

[jknezek]

The most over-rated team is: E&H.

The complete numbers are here:
W&L   0.625
Catholic   0.5
Guilford   0.375
RMC   0.25
HSC   -0.125
Bwater   -0.5
E&H   -1

Interesting. Keith and I often have discussions that center around, "are we overvaluing our alma maters" -- especially in years where they are further down the table. This suggests we're not doing that, at least.

This means even as we overrate E&H or underrate W&L, it's been by an average of one spot per year? I'm OK with that.

Sort of. That's why I started out with the fact that it is statistically insignificant. One important thing to remember is that I didn't measure how much you missed by in absolute terms, I measured by how much you  under- or over-valued on an average basis. In other words, I was searching for a consistent bias, which I don't think I found in any signficant amount. However, if you wanted to know which teams you average the highest absolute misses on, the data would look like the following:

Bwater   1.5
HSC   1.5
RMC   1.75
Catholic   1.75
Guilford   1.75
W&L   2.5
E&H   3.5

So you miss the least in Bridgewater, despite consistently missing high, and HSC and miss the most on W&L and E&H. Considering who you miss the most on, and the other analysis provided above, you would have been significantly more accurate by polling W&L a spot or two higher every year and E&H a spot or two lower. However, the problem with E&H was basically in the first years of the data, the problem with W&L is mostly consistent, but exacerbated by the outlier years of 2008 and 2010.

There is one other way to look at the data that may be somewhat interesting, though I'll repeat that I don't think any of this is statistically significant. However, there are 8 years data available. In those 8 years, here are the number of years a team is over or under valued (if they don't equal 8, it's because the remainder was spot on).

   Over   Under
Bwater   4   2
HSC   3   3
W&L   2   5
E&H   5   2
RMC   2   3
Catholic   2   3
Guilford   3   3

What stands out? Bridgewater and E&H have the most years over, W&L has the most years under. W&L and E&H both were only correctly picked once. What accounts for this? There probably is a bit of bias in the rankers. For example, E&H was overpredicted in the earlier years, which could be accounted for by the fact that they were really good, or expected to be, when the rankers were in school (I think. I believe you and Keith are about my age, which means school in the mid 90s to early 2000s roughly). Remembering that could have boosted E&H in the early years. Similarly the data starts just as Bridgewater tailed off their run. They were picked first in the ODAC in each of the first 3 years of data, but finished 1,2, and 4. The Eagles were overvalued or spot on in 5 of the first 6 years.

W&L was not real good at the time we were in school. Not putrid, but no better than mid-conference. The 2006 team wasn't much of a surprise, picked 2 finished 1, but that 2 was W&L's only real non mid-conference pick until the 2011 team, and even the 10 and 12 championship teams weren't all that respected preseason at 6 and 3. W&L has finished 3 or better 6 of the 8 years, but were picked 3 or lower 5 of 8 years. So this year's 5 is, most likely, another under ranking. Especially since Guilford and Bridgewater both have to play in Lexington.

But since W&L didn't start picking up until after the data starts, it is understandable that they were consistently undervalued. This year will be telling. If they finish 3 or higher, there probably is an issue with the rankers perception of W&L now, as opposed to how they remember them. If W&L finishes 4 or lower, it's possible that the "good years" of Westfal and Heinsohn were an outlier that was tough to predict and will be hard to match.

Then again, it's a small sample, so it really is what you make of it.

[Pat Coleman]

I don't think it's related to the 90s but two factors. First of all, we expected Don Montgomery to have a positive impact on that program, which has not materialized, and secondly, W&L's performance leading into the "Kickoff era" of 2005:

2004 (6-4, 4-2 ODAC)
2003 (3-7, 1-5 ODAC)
2002 (5-5, 3-3 ODAC)
2001 (5-5, 3-3 ODAC)
2000 (5-5, 2-4 ODAC)
1999 (5-5, 2-3 ODAC)

I graduated in 94 and I think Keith graduated in 1997. I don't really know what those teams looked like.

Our ODAC writer attended W&L and I don't think we made any real adjustment on his picks this year. (Although ODAC funnels up through Keith for editing and not me, so I can't say for sure.) I think we all try to be as fair and unbiased as possible. Perhaps we are being too much so.

[Pat Coleman]

I'd be interesting in finding out how the preseason coaches' poll does in that kind of comparison as well. The ODAC has been fairly volatile, especially since Bridgewater stopped dominating the top spot.

[jknezek]

I don't think it's related to the 90s but two factors. First of all, we expected Don Montgomery to have a positive impact on that program, which has not materialized, and secondly, W&L's performance leading into the "Kickoff era" of 2005:

2004 (6-4, 4-2 ODAC)
2003 (3-7, 1-5 ODAC)
2002 (5-5, 3-3 ODAC)
2001 (5-5, 3-3 ODAC)
2000 (5-5, 2-4 ODAC)
1999 (5-5, 2-3 ODAC)

I graduated in 94 and I think Keith graduated in 1997. I don't really know what those teams looked like.

Our ODAC writer attended W&L and I don't think we made any real adjustment on his picks this year. (Although ODAC funnels up through Keith for editing and not me, so I can't say for sure.) I think we all try to be as fair and unbiased as possible. Perhaps we are being too much so.

I don't think there is anything unfair or biased about it. It's a darn good attempt at the impossible. What I posted wasn't so much a critique as a bit of fun. There is nowhere near enough data to draw anything solid, and the numbers are close enough that anyone should consider them really darn good. I would be truly amazed if the 2011 accuracy repeated itself with any real accuracy given the information and communication challenges inherent in D3. Heck, even in D1 or the pros, where information is so much more readily available, it's darn hard to predict rankings the way Kickoff goes at it. Otherwise Vegas would be in much more trouble.

My analysis was just a bit of fun as I think W&L is a spot or two low this year and I think Bridgewater is high. I was curious if that is consistent. Imagine my surprise when I discovered that, based on limited data, it might be consistent. But I knew going in the data wasn't going to provide anything near an actual answer and, of course, it didn't.

Regardless, good fun and thanks for pulling it all together. Now bring on some actual games and we'll see whose preseason guesses come true!

Let me see if I can find the coaches poll historically.

[Pat Coleman]

I'm not taking it as a critique, or anything unfair -- sorry if I come off that way. Definitely interested.

[jknezek]

The coaches poll info isn't real good. I found 5 seasons worth, 2007-2008, 2010-2012. 2009 is not listed on the ODAC website like all the others.

The data just isn't a good comparison thanks to the missing seasons, but here it is:
D3 average bias                             Coaches average bias
W&L   0.625                 RMC   1.4
Catholic   0.5                 Catholic   0.8
Guilford   0.375                 W&L   0.2
RMC   0.25                 Guilford   -0.2
HSC   -0.125                 HSC   -0.2
Bwater   -0.5                 E&H   -0.8
E&H   -1                 Bwater   -1.2

D3 Absolute Miss *Corrected*        Coaches Absolute Miss
Bwater   1                              Guilford   0.625
Catholic   1.25                 Catholic   0.75
Guilford   1.375                 RMC   0.875
HSC   1.375                 Bwater   1
RMC   1.5                 E&H   1.25
W&L   1.875                 HSC   1.375
E&H   2.5                 W&L   1.375

D3    Over Under   (8 seasons)        Coaches Over Under      (5 Seasons)
   Over   Under         Over   Under
Bwater   4   2      Bwater   4   1
HSC   3   3      HSC   3   2
W&L   2   5      W&L   3   2
E&H   5   2      E&H   3   1
RMC   2   3      RMC   0   3
Catholic   2   3      Catholic   2   1
Guilford   3   3      Guilford   2   1


Couple thoughts on this, it looks like RMC is severely under-valued, but that is pretty much all tied up in a single outlier. It is the biggest outlier in the 5 years of the coaches poll, and otherwise the coaches were no more than +1 in the other 4 years. So that big outlier really skews the data over only 5 points. Bridgewater, however, is still over-valued. 4 of 5 seasons they are at least a -1, with only the 2012 poll coming up with an undervalue of +1. So year over year, Bridgewater gets the benefit of the doubt.

W&L appears to be less undervalued by the coaches, although still one of three teams undervalued. Catholic, much like RMC, also has a single outlier, a +4 in 2008 when nothing else is more than +/-1.

Why is the data so different? The missing early years as Bridgwater came off their run, were hard years for D3 to predict. There is no knowing if the coaches would have done any better.

With so much data missing, I don't think this is anything near a good comparison.

[tigerfanalso]

Jk

You have way too much time on your hands but I did enjoy your post.
I don't see the ODAC being as competitive as advertised this year. Will not go into detail but I think two teams will fight it out
for the championship and those two play each other on Nov. 16th.

[jknezek]

Jk

You have way too much time on your hands but I did enjoy your post.
I don't see the ODAC being as competitive as advertised this year. Will not go into detail but I think two teams will fight it out
for the championship and those two play each other on Nov. 16th.

That is my thought as well. W&L will not beat both RMC and HSC on the road. Bridgewater won't beat W&L and HSC on the road. Guilford is too young and the QB isn't going to sneak up on anyone this year. HSC will run the table until The Game. RMC will struggle at Bridgewater the week before The Game. Who comes out from that tangled mess? I think 1 loss RMC (Bridgewater) and HSC (losing to RMC) will play for the title. A 1 loss W&L team (at RMC) will hang around until HSC hangs a 14 point loss on them. Bridgewater's joke of a first four games gets them smacked around at HSC and again in Lexington.

RMC 1 loss champ, 2 losses total. 7 seed
HSC 1 conf loss in the last week, 5 or 6 seed pool C maybe
W&L 2 total losses, 1 conf loss
Bridgewater 2 total losses, 2 conf losses