WrestleStat Rankings - Week 16 2018-2019

[Andegre]

Rankings have been updated, but I need to point out some changes that occurred last night.

1) Wrestlers records have been updated to reflect only Division I opponents. This was done to fall in line with the same criteria as the NCAA's RPI rankings use for the 17 match minimum.
2) Wrestlers profile pages have been updated to display both the [Division I] record, as well as the wrestlers overall record. This will show up in the Season Summaries section. The overall record will be in parenthesis. Check out Chance Marsteller's profile for an example: https://www.wrestlestat.com/wrestler/47805/marsteller-chance/profile
3) Also, on a wrestlers profile page is now the RPI. This is the last column in the Season Summaries section. Right now it's displaying the actual RPI value, but that may change or get removed since it doesn't provide much context relative to other wrestlers. Anyway, since the official RPI is based on the 17 match minimum, not everyone will have an RPI. In order to get a guage about where, approximately, a wrestlers will fall once they reach that 17 match minimum, there's a 2nd RPI value in parenthesis. This RPI is based on a 5 match minimum. For example, Stevan Micic at 133 (ranked #1) has only had 13 matches so far this year, therefore does not have an official RPI (so it shows NA), but it is showing the unofficial RPI so you can get an approximate idea of where he will land once reaching that 17 match minimum.
4) Lastly, since we're on the subject of the RPI, I went ahead and added an RPI rankings page. This page is initially sorted by the UNOFFICIAL RPI (smaller number at the far-left), but will display in bigger font the official RPI. Again, this is to give an idea where a wrestler will slot-in once they reach the 17 match minimum.

Keep in mind, I'm not 100% sure on the criteria of the RPI calculations that the NCAA uses, but I'll continue to tweak once I find out all of the details. As of right now, it's calculated like this: (25% Wrestlers Win %) + (50% Opponents Win %) + (25% Opponents Opponents Win %)

Now, for the rankings...I've wrote too much already, so here they are:

Wrestler Rankings: https://www.wrestlestat.com/rankings/starters

Dual Rankings: https://www.wrestlestat.com/rankings/dual

Tournament Rankings: https://www.wrestlestat.com/rankings/tournament

Statistical Rankings: https://www.wrestlestat.com/rankings/statistical

Redshirt Rankings: https://www.wrestlestat.com/rankings/redshirts

RPI Rankings: https://www.wrestlestat.com/rankings/rpi

 

[BillyClone]

I know you get flak about your rankings but I have always liked your site because of all the data you provide. These changes are definitely an improvement and I really appreciate you guys gathering this information. I don't know of another site where I can find Div-1 only records. Thanks!

 

[Andegre]

Hey @GrappleCy I got the RPI's added to the site using the info you provided last week, thank you!

Now, a question, any idea how/why Austin Gomez is #1 at 133 while the NCAA RPI officially has him at #5? Obviously mine is updated after match results are entered, but still seems like quite the difference.

 

[GrappleCy]

Hey @GrappleCy I got the RPI's added to the site using the info you provided last week, thank you!

Now, a question, any idea how/why Austin Gomez is #1 at 133 while the NCAA RPI officially has him at #5? Obviously mine is updated after match results are entered, but still seems like quite the difference.

The released RPI is a week old at this point but there weren't that many matches since then. Once you get in to the really gritty details it is tough to say since the documentation that's released tends to be incomplete and conflicting - the NCAA website itself gives two different ways to calculate RPI.

Here's my first guess though - You don't need to be eligible for an RPI to have your win percentage included in someone else's RPI calculation. However, you do need 8 matches to have a win percentage calculated for the win percentage leg of allocations. Maybe you also need 8 D-1 matches to have your win percentage included in RPI? The only other thing I could think of is that it is something to do with injury defaults vs. forfeits vs MFFs.

EWdit: Here's the NCAA site with all the documents they have for D-1 Wrestling. I can't find anything else that'd help but maybe you'll have better luck.

 

[Andegre]

The released RPI is a week old at this point but there weren't that many matches since then. Once you get in to the really gritty details it is tough to say since the documentation that's released tends to be incomplete and conflicting - the NCAA website itself gives two different ways to calculate RPI.

Here's my first guess though - You don't need to be eligible for an RPI to have your win percentage included in someone else's RPI calculation. However, you do need 8 matches to have a win percentage calculated for the win percentage leg of allocations. Maybe you also need 8 D-1 matches to have your win percentage included in RPI? The only other thing I could think of is that it is something to do with injury defaults vs. forfeits vs MFFs.

Thanks, and I'm guessing they'll never release the full details of the calculation, so best I can do is go by the generic one that I implemented. Should be "close enough"...

Thanks again!

 

[GrappleCy]

Thanks, and I'm guessing they'll never release the full details of the calculation, so best I can do is go by the generic one that I implemented. Should be "close enough"...

Thanks again!

Honestly you might have luck if you reach out to Iowa's SID and ask if they can get you the documents. I've reached out to SIDs before for clarification on things (Mainly when I was salty about Gadsen's bad seed his junior year) and they were able to track stuff down for me. Most of the time the NCAA isn't trying to hide stuff, they just don't do a good job with getting it out there if you don't already know where to look.

 

[cyclonegrap]

Thanks, and I'm guessing they'll never release the full details of the calculation, so best I can do is go by the generic one that I implemented. Should be "close enough"...

Thanks again!

I love the new changes. I am confused on the two RPI scores. Shouldn't these two numbers be the same if someone has over 17 D1 matches? Also, winning percentage for RPI is only calculated on matches at the same weight. For example Jacobe Smith has a 22-3 record vs D1 opponents, but has a 13-3 record at 184 and 9-0 record at 174. The 13-3 record is used for RPI winning percentage when he wrestles someone at 184 and the 9-0 record is used when he wrestles someone at 174. He is one match away from being RPI eligible.

 

[buf87]

@Andegre,
so is the record that is listed on the RPI ranking page, the overall record or the D1 record that is used for the RPI ranking?

 

[Andegre]

@Andegre,
so is the record that is listed on the RPI ranking page, the overall record or the D1 record that is used for the RPI ranking?

Everything on the site, EXCEPT for the record in parenthesis in the season summary section of a wrestlers profile page, is going to be the D1-only record.

 

[Andegre]

I love the new changes. I am confused on the two RPI scores. Shouldn't these two numbers be the same if someone has over 17 D1 matches? Also, winning percentage for RPI is only calculated on matches at the same weight. For example Jacobe Smith has a 22-3 record vs D1 opponents, but has a 13-3 record at 184 and 9-0 record at 174. The 13-3 record is used for RPI winning percentage when he wrestles someone at 184 and the 9-0 record is used when he wrestles someone at 174. He is one match away from being RPI eligible.

No, the "unofficial" version is only requiring 5 D1 matches, therefore it's a larger body of wrestlers that have their win % included in the RPI calculations.

 

[cyclonegrap]

@Andegre

Another question on how you calculated the RPI. I was trying to calculate the RPI for a single wrestler to compare it to your numbers. It turns out that it is a lot of work.

The formula is (25% Wrestlers Win %) + (50% Opponents Win %) + (25% Opponents Opponents Win %), but it seems like the devil is in the details.

To calculate the Opponents Winning Percentage (OWP) do you take the weighted average (all the opponents wins divided by the total number of matches) or do you take an average of the winning percentages? It look like NCAA basketball does the latter.

For example, if a wrestler A has a 3-2 record (Wins against B,C,D and Losses against E,F).

Wrestler B is 0-6 (0.0)
Wrestler C is 0-6 (0.0)
Wrestler D is 2-4 (0.33)
Wrestler E is 21-0 (1.0)
Wrestler F is 4-2 (.67)

First, you have to remove the matches between the wrestler in question and all the opponents so the adjusted records are:

Wrestler B is 0-5 (0.0)
Wrestler C is 0-5 (0.0)
Wrestler D is 2-3 (.40)
Wrestler E is 20-0 (1.0)
Wrestler F is 3-2 (.60)

I would have thought the OWP would be Wins/(Total matches) = 25/40 = 0.625
It looks like basketball RPI takes the averages of the winning percentages (0.0 + 0.0 + 0.4 + 1.0 + .6) / 5 = 2/5 = 0.4

Just curious which method you used. Thank you.

 

[Andegre]

I do the latter as well (average of opponents winning percentage instead of adding up wins and losses, etc).

But, I I don't believe I have the formula quite right yet...I'm getting clarification, and will probably update it again this weekend.