re: The Big 10 and ACC are setting the stage for locking up 3/4 playoff spots

quote:

---

JayWhite

---

I found the contrarian...who thinks he has a hot take...

quote:

---

Well I think maximizing objectivity should be a goal, and things like Massey are currently pretty great for that and have done a great job so I’m trying to not throw the baby out with the bath water.

---

We tried that, and it got us right back to my point all along. Massey and Sagarin support what I've been saying.

quote:

---

The creation of a minimum game threshold until we actually trust the Massey Composite doesn’t seem that unreasonable to me and I’d think we could track how much volatility the composite has week to week over the previous 20 seasons and once the volatility shrinks sufficiently in each season we can understand how many games are needed until the rankings are stable for your average season. There is no way to eliminate subjectivity here, but it seems like a pathway to minimize it.

---

Minimum game requirements are inherently arbitrary. There's no way we could objectively say what they should be from year to year, and certainly no way to have done it prior to this season.

You admitted that the best option we currently have for objectivity mostly supports the current rankings.

If your point is simply that we should continue to improve the process, sure, I'd agree with that.

If it's that we have some better way of determining the best four teams this year, I don't think I'm convinced.

quote:

---

We tried that, and it got us right back to my point all along. Massey and Sagarin support what I've been saying.

---

I said “akin to Massey” originally and for this reason. They are the best composite of metrics that we have, but we can strive to improve them and I think 2020 is exposing a weakness in the metrics.

quote:

---

There's no way we could objectively say what they should be from year to year, and certainly no way to have done it prior to this season.

---

If you analyze decades and decades of data regarding week to week team ranking volatility, and take the a standard deviation less than the mean number of weeks towards 90% convergence to keep it on the less restrictive side; or some type of exercise similar to this, I don’t see how that’s super unreasonable or excessively subjective.

They should have just done a 6 or 8 team playoff for one year and let the conferences send their champion and let in a couple of group of 5 or non conference champions.

Since we are just making rules up as we go why not.

quote:

---

I said “akin to Massey” originally and for this reason. They are the best composite of metrics that we have, but we can strive to improve them and I think 2020 is exposing a weakness in the metrics.

---

You sure did. And the best composite of metrics we have still agree with what I've said.

Sure, let's improve them. What does that look like? If it's a random insertion of minimum game requirements, I'll pass on that.

quote:

---

If you analyze decades and decades of data regarding week to week team ranking volatility, and take the a standard deviation less than the mean number of weeks towards 90% convergence to keep it on the less restrictive side; or some type of exercise similar to this, I don’t see how that’s super unreasonable or excessively subjective.

---

What do the last few decades of data say about teams like Cincinnati going up against Bama, Clemson, Ohio State, etc? What do they say about about teams who were thoroughly beaten by four touchdowns? What about Ohio State versus their conference.

If we're going to take in data to fill in the gaps, let's do that, but we can't ignore what's inconvenient to our desired conclusions.

quote:

---

A&M deserves that 4 spot. No fricking debate!!

---

quote:

---

the 12th spot for a team that’s played 12 is often against an opponent who might as well be a bye for a team trying to make the CFP

---

For the reasons in the original quote and a few others, I really do not hop on board with this stance entirely, although I get what you are driving at.

I’d also say the marginal utility between an eleventh and twelfth sample is dramatically different than between a hypothetical sixth and tenth sample, rendering the original argument I put forth less relevant

quote:

---

If we're going to take in data to fill in the gaps, let's do that, but we can't ignore what's inconvenient to our desired conclusions.

---

Again, no skin in this game and I completely get this attitude. I’m an engineer by trade and this is something that comes up often with numerical simulations. You should be able to demonstrate predictive power with a numerical method for a given experiment without having to alter the methods between runs.

However, I’m trying to address what I perceive to be a glaring weakness in using our quasi-objective metrics with an improvement. You don’t seem won over by my idea of an improvement, and frankly it won’t matter past this season as I doubt we have to cross this bridge again, but I do think we need a fair and quasi-objective way to normalize between dramatically different sample sizes and I’m not sure any algorithm used in Massey really does that at all

quote:

---

OSU is not getting in without playing this weekend.

---

Their whores on the CFP committee will still let them in.

quote:

---

However, I’m trying to address what I perceive to be a glaring weakness in using our quasi-objective metrics with an improvement. You don’t seem won over by my idea of an improvement, and frankly it won’t matter past this season as I doubt we have to cross this bridge again, but I do think we need a fair and quasi-objective way to normalize between dramatically different sample sizes and I’m not sure any algorithm used in Massey really does that at all

---

What's a bit amusing is that the SEC has still been playing an 8 game conference season in hte past. It's a clear advantage vs conferences that are now playing 9 conference opponents in a normal season. There are inequities between conferences in scheduling and it would be nice to see a baseline established to even out the playing field.