Posted by gillispie1 on 4/7/2016 9:56:00 PM (view original):
Posted by tarvolon on 4/7/2016 8:09:00 PM (view original):
Posted by gillispie1 on 4/7/2016 7:37:00 PM (view original):
my internet crapped out so it took a long time for the edit i put in right after my post to make it through, but what i said applies... how many final fours did you make without top rpis? the event success according to NT given the event of having a top rpi already, cannot be used to indicate that you do worse in the NT than by rpi. you have to compared the event of success according to the NT to the event of success according to RPI...
Top 2 RPIs that didn't result in Final Fours: 16
Final fours without top 2 RPIs: 10
Of course, every season in every world has at least two final four teams without top 2 RPIs. So. . .
Final fours without top 4 RPIs: 4
Final fours without top 6 RPIs: 2
#choker #orpossiblyoutschedulingmyself
16 vs 10, thats pretty close!
its not exactly fair to compare rpi with post season to post season as if they are independent, as you said earlier. those 10 final fours without top 2 rpi, would all still be lower than 2 rpi, if you looked at regular season rpi, which is the independent ranking. but with post season, your rpi goes up, so you are getting final 4s in there where you had less-than-top 2 rpi going into the thing, and top 2 coming out - so its actually closer than 16 v 10, if you look at that.
scheduling lets you inflate rpi, and most coaches with technical mastery schedule well and inflate their rpi above their quality of team. so a good measure would be comparing something like regular season rpi vs final four, only using top 2 rpi, to compensate for the rpi inflation you are doing. so, you are looking at a more fair than 16 v 10 comparison, in that kind of light - probably significantly more fair - and 16 v 10 isnt that far off to start. so, your performance on the two lists, by that measure, is probably very similar.
so, like i said in the beginning - making something of nothing :)
This is actually not quite right, for a few reasons.
(1) You're right that they aren't independent, but you can't use the lack of independence as a reason to hold the 16 number steady and move the 10 up. Yes, I probably made several final fours where I didn't start in the top two but finished in the top two after the Final Four run. So the 10 number probably goes up. But if you're using pre-tournament RPI and not post-tournament RPI, then the 16 number has to change too. And there are surely cases where I started with an RPI in the top two, lost early, and then finished outside the top two. When you try to adjust for independence with one number and leave the other fixed, you're comparing apples with oranges. I know I'm using an RPI number that isn't independent of tournament success, but I'm doing so consistently.
(2) You've pointed out that 16 and 10 are pretty close. But you haven't asked whether we should expect them to be close. When you compare the categories of "top two seasons without a final four" and "non-top two seasons with a final four," it's important to remember that, over the course of WIS history, the latter category has
2661 more teams in it than the former. So we should expect the latter category to be much bigger, not pretty close as big. If you want to compare categories that are the same size, you should compare RPI top four seasons without a Final Four and non-RPI top four seasons with a final four. Personally, I have had 25 teams that finished in the RPI top four without a Final Four and 4 teams that finished outside the RPI top four and did make the final four.
25 to 4 is a pretty big difference. Like, a really huge difference. As you pointed out, smart scheduling can inflate your RPI, and I do consider scheduling to be something I'm pretty good at (if you want proof, just look at the RPI of my Dallas team right now--it's absurd! Actually, my Dallas team last year had a top ten RPI after non-con and didn't even make the tournament. So yeah, I'm good at inflating RPIs, probably one of the best at D2 and D3). And so RPI inflation probably accounts for a good chunk of that difference.
To be perfectly honest, I didn't think the Lemoyne-Owen team that started this rant was the best team in the country going into the tournament. I thought it was a solid team whose RPI was inflated by smart scheduling. But after the tournament played out, I saw that LMO's loss was basically getting RNG'd (allowed 54% shooting to a team that wasn't exceptional at shooting, despite LMO winning basically on the back of FG% defense all season), and that LMO was 3-0 against the teams in the title game (and 2-0 against the champ), and I started to think maybe this team was better than I thought and their one bad sim just came at the worst time. And that's kinda how I feel here. I know scheduling is a factor. But 25 to 4 is a pretty ridiculous gap. Honestly, I don't think RPI inflation accounts for all of it. I've had too many early exits with teams I thought were clear title favorites and too few deep runs with teams I thought were solid but nothing special.