This offseason is taking forever. Luckily for me, that means I have more “time” to work on enhancements to the website. If you don’t follow wrestling-related message boards, especially during the offseason, then you probably aren’t aware of the changes that are going on right now, and there’s A LOT!
This, in my mind, is the biggest (and coolest) enhancement of all. Now that WrestleStat has 3 seasons of data to process, it’s starting to get “career” results for all wrestlers. With that data, I wrote an algorithm to try and “predict” the outcome of any particular match. For example, Thomas Gilman vs Nathan Tomasello shows a prediction of 6 – 4, in favor of Gilman. (Click here to see this prediction). This will be a very fun, and handy tool come next season to give fans another look, besides just common opponents, as to who might come out ahead. Since this is the first year of this new process, it may have some skewed results for some matches. There will definitely be tweaks made in the future, but the base is already there. One thing to note with this prediction, for wrestlers that have never wrestled a college match, there’s really no data to go on besides their default ranking. So even though we know someone is going to be good, like Mike Hall or Spencer Lee, until they wrestle a college D1 match, they will be treated the same as any other wrestler with a default ranking.
Rankings – Individual
There are 2 significant changes going on with the rankings processes. The first one has to do with how individual wrestlers rankings are generated. Since I’m no mathematician, implementing the Elo ranking system into WrestleStat is been a VERY tricky process. The first year, there were numerous tweaks here-and-there in order to try to make the algorithm work for as many different scenarios as possible. But like most ranking/rating processes, there will always be outliers that just can’t get accounted for.
Thankfully, there happens to be a frequenter of the website that is also an engineer and really into math. He has provided me with additional tweaks to the ranking algorithm to try and get it more accurate, and also reduce the amount of outliers that pop up on a weekly basis.
I’m still in the process of implementing his changes, as well as a few more tweaks that are needed in order to make it work better season-long. I hope to have this change finalized prior to the new season because I don’t want to change the algorithm mid-season.
Rankings – Dual Rankings
The second enhancement in the rankings process has to do with how dual rankings are generated. In seasons past, the dual rankings were generated/based on taking the sum of each of the 10 individual starters. That total was then used to rank all of the teams from 1 – 75/76. Click here to see the current dual rankings.
The new version of the dual rankings will be implemented utilizing the prediction engine, as described above. In this process, it will cycle through all 76 D1 teams, and run them through the prediction engine vs the other 75 teams. At the end of this process, the ranking will be determined by the schools number of wins. For example, Penn State could end up with a record of 73-2, or Bloomsburg could end up with a record of 38-37. If there are two teams that have the same number of wins at the end of this process, then the old-style team rankings process will serve as the tie-breaker (getting the sum of all 10 starters rankings).
Complete WrestleStat Re-Write
Along with those other changes and enhancements that have been mentioned already, the last thing to note is that the ENTIRE website is getting a complete overhaul, in both looks, as well as the underlying code behind the scenes. There are multiple reasons why I’m doing this, but the major ones are:
- Old look-and-feel I didn’t like
- Ease of maintenance
- Ease of implementing new enhancements
- Allow me to reduce the amount of time I spend on the site so I can spend more time with my family
- Allow for the possibility of expanding beyond just Division I wrestling
I have spent way more time working on this website than I even want to consider. It’s taking a toll on me [and my sleep]. Just last night, I worked from 10:00pm until 3:30am. Do that a couple times a week, and that leads to other areas of my life suffering.
That’s a high-level view of the changes that are going on right now. If you’d like to check out the new design of the website (if you didn’t click any of the links above), you can click here and check it out. This is still a work-in-progress, but the framework is pretty close to established, I just still need to add lots of pages and functionality from the old site.
As always, I appreciate any and all feedback, both positive and negative. If you have any comments or suggestions, feel free to leave a comment, or send me an email at firstname.lastname@example.org.