First NCAA NET rankings released (they are pretty absurd)
- Thread starter JFCats22
- Start date
Insult to injury about the Duke loss is that margin of victory/loss actually is factored in and I can already see the Committee Chairman justifying a seed for UK..."well they won every game this season, but they lost by 34 points to Duke 5 months ago".
I think it's really only gonna come into play is if both Duke and UK were fighting for the same seeding in which case they can say look at what Duke did to them head to head.
Otherwise it is what it is.......1 game in a 30+ game season.
Honestly, delete your account if you buy this.
People are automatically buying TennSt and Monmouths next-to-bottom-ranking in a metric that actually thinks Kentucky isn't a top50 team. A total crock, especially considering we still (?) don't know what the criteria is.
I don't believe in the Ignore button, but it is really tempting lately.
People are automatically buying TennSt and Monmouths next-to-bottom-ranking in a metric that actually thinks Kentucky isn't a top50 team. A total crock, especially considering we still (?) don't know what the criteria is.
I don't believe in the Ignore button, but it is really tempting lately.
I don't think anyone believes that UK is the 61st best team in the country. If they do, I dunno what to tell them lol.
That being said, the NCAA is going to use these. They replace RPI. To what extent they are used for seeding remains to be seen, but I do think in the future when they do stabilize, it'll be important to look at.
Also, how about this "If you take the Duke game away, we played no one good!" argument. Wut?
Top 25 teams that haven't beaten a decent opponent:
Nevada
UCLA
Purdue
Nova (Beat FSU but lost to Furman, push)
Maryland
Miss St.
We're 6-7 games into the season. Taking away any one game from any one team changes their outlook drastically.
Top 25 teams that haven't beaten a decent opponent:
Nevada
UCLA
Purdue
Nova (Beat FSU but lost to Furman, push)
Maryland
Miss St.
We're 6-7 games into the season. Taking away any one game from any one team changes their outlook drastically.
I have no doubt it will stabilize. The Duke beating must have weighed heavily on this. It's certainly weighing heavily on some "fans". That game is more and more looking like an unfortunate mismatch. I don't think any single other team would have beaten Kentucky by 15+ on opening night, despite how bad we looked.
lol Nate Silver said this is the worst ranking system he has ever seen........in any sport lol.
Honestly, if the NCAA wanted a decent ranking system they could have just used Kenpom or Sagarin lol
There was an article on it several months ago but I haven't read much about it.
I know it factors in margin of victory, sos, offensive and defensive efficiency......but I dunno. I'm curious to find out tho.
Whatevs. While I would have liked a bit tougher opponents to start, I'm extremely glad we have a schedule that affords us the ability to go 9-1, while holding the tougher games until mid-december and on.
If we're as bad as some fans think, imagine if we put Kansas, UNC, Louisville and one of SH/Utah in the opening weeks.. we'd be 2-5 and all hell would break loose.
Personally, I hate the idea of playing a ranked team in the first month or two, as long as we have freshman and sophomores. Playing tough teams early on is a disservice to Kentucky, who would probably beat those same teams later in the season. Why should we take avoidable losses early on?
I agree. I don't mind the schedule like this.
Tho I would tell Cal schedule more teams in the 150-200 range than the 250+ range.
These games are at Rupp. We'd still win them and that way it doesn't completely kill our SOS. Our non conference is always overall good because we play some big teams but it's very top heavy. Just a minor issue tho.
Tho I would tell Cal schedule more teams in the 150-200 range than the 250+ range.
These games are at Rupp. We'd still win them and that way it doesn't completely kill our SOS. Our non conference is always overall good because we play some big teams but it's very top heavy. Just a minor issue tho.
Agreed 100%.
First 6 games: 200th-100th
next 6 games: 150th-1st (even mix).
It's really doing a disservice to Kentucky to be playing tough games when our front court is still getting to know our backcourt.
Exactly. It's not that we've played mid-majors, instead it's that we've played the very bottom rung of mid-majors. There's a huge difference between playing say a MAC team and playing VMI, Winthrop and Monmouth.
Beating that bottom rung does nothing for your computer ranking.
Last edited:
The problem with that is that you're using their awful metrics of ranking teams to determine what a good or bad win is to justify their other awful rankings.
Saying we don't have a good win because the NET ranks them low when the data is clearly a disaster, especially this early on, is pointless. It's begging the question.
We have what might be the best OOC schedule in the country this year.
I read that the margin of victory factor was limited to 10 points. So if team A beats team B by 10 at home, it's worth the same as Team C beating team B by 40 at home.
I know the NCAA has always been scared of margin of victory, but they would be much better off accepting the reality that any decent system is going to factor for it, and any decent system will have limitations on it already built in. Coming up with some artificial cut-off is stupid (no surprise), and is probably a big part of what caused the absurd initial rankings.
Won't mean anything until the end of December, anyway, but if it's still absurd then, watch the NCAA back off of it. Despite the fact that it would be easy to use a variety of computerized systems to come up with a consensus for seeding (and to go with it and eliminate the human element), the NCAA just can't resist clouding things in unnecessary mystery and playing back-room games.
This.
Seems like something that would have been better suited being released no earlier than the beginning of conference play.
Probably could have waited until Feb 1 and made it similar to the CFP committee rankings.
Yeah 10 is ridiculous.
I like how Kenpom does it where you get to a point and it's diminishing returns. In other words the difference between a 10 and 20 point victory is greater than the difference between a 20 and 30 point victory. But even in that system even with the diminishing returns a 30 point victory still means more. To stop at 10 is just silly.
That's probably how it should be done.
And here's my thing.......
You've got a TON of good system out there.
There is really no reason to take a bad system (RPI) and replace it with another bad system (NET).
Wisdom of the crowds really works well......take all the highly respectable systems.......average them.......and use that. It would probably be one of the most accurate systems.
You've got a TON of good system out there.
There is really no reason to take a bad system (RPI) and replace it with another bad system (NET).
Wisdom of the crowds really works well......take all the highly respectable systems.......average them.......and use that. It would probably be one of the most accurate systems.
Kansas has beaten two top ten teams on a neutral court and isn’t even in the top ten 
I get why Kentucky isn't on here.
I get why the Top 10 teams are there and a computer this early in the season might fort them differently than I would.
But why on earth is Ohio State #1.
I get why the Top 10 teams are there and a computer this early in the season might fort them differently than I would.
But why on earth is Ohio State #1.
Kansas has beaten two top ten teams on a neutral court and isn’t even in the top ten
BUT WE'RE LOW CAUSE WE AINT PLAYED NOBODY PAWWLL.
Their system is pretty awful, and it being this early accentuates all that awfulness.
So when you have a team that's played one more game than another to be 6-0 instead of 5-0, that's going to factor too heavily.
or when a team like Ohio State has 2 road wins where most teams that are unbeaten have only played home/neutral games, those road wins are going to factor heavily and throw things off.
When nearly all of their wins are by 10+ points because they've played relatively weak competition (like most everyone else) that is going to factor too heavily.
Three scoring outputs of 87 points or more while holding opponents to no more than 62 makes their efficiency and margin numbers look good.
It's just way too early, especially for such a weird system.
All metrics get more accurate with more results. Hopefully this happens for the new system.
The margin of victory is potentially a problem, but there are others. Using raw rather than weighted efficiency data, for starters. Could be balanced out by the Team Value Index, but that’s hard to know.
Honestly, I would question their methodology based on the write up at the NCAA site. They talk about using late season games in their test sets - did they use data from games to predict the results of those same games? Serious issue if so.
I also get that machine learning is all the buzz, but you need a whole lot of data to make a machine learning model useful. There are academics who have struggled to build predictive machine learning models for the NCAA Tournament for this reason.
Put simply, if you’re trying to predict tournament success following one season, there may not be enough games in a season for a machine learning model to do a quality job. Simpler models are often better. I’m not saying it can’t be done, but I would really like to see the methodology and results, rather than take the NCAA’s word for it.
Do like the football rankings and wait until Jan 1, after most non-conference games.
Yep. Not only should we be going for a #1 seed but THE OVERALL #1 SEED. Every loss and poor performance gives the NCAA and their ridiculous selection committee another excuse to seed us poorly. It happens just about every year.
The only way Kentucky is not in the same bracket as Duke is if Kentucky gets a 1 seed. Duke will be a 1. The selection committee is already dying to put us as their 4 or 5, or their 8 or 9 of we once again struggle on the road in conference this season.
Yeah I don't know too much about machine learning but using the same data to train and then test a model is a big no no. The model suffers from overfitting and then when new data is introduced, it's horrible at predicting.
Picking where a team will end up the first two rounds isn't difficult but picking a region I think is because it doesn't just matter what seed you are but also where you are on that seed line.
There's a big difference between being the 1st 2 seed of the last 2 seed. The last 2 seed, we probably end up out West with Gonzaga. The first 2 seed, MAYBE we end up in Louisville. Tho I highly doubt they'd give us that geographical advantage over whoever is the 1 seed in that region.
Tho we are getting ahead of ourselves lol
Cal keeps saying “it’s early don’t care”. This is not the mind set of a coach trying to win titles. It’s the mindset of a coach who is trying to develop a team. It would be fine if it wasn’t the same “it’s early” every year. He needs to start caring about what happens in November and December bc the selection committee does.
No college coach is good enough to develop freshmen over the course of a year then win the ncaat from the 3,4,5+ line.
Similar threads
