Special Episode – Ken Massey from Massey Ratings

On today’s episode, I speak with Ken Massey from Massey Ratings. Ken has developed one of the most unique and informative ranking systems for all sports, including college soccer. Ken and I discuss his formula and how it can help those looking at comparing programs. Learn more about Ken Massey from Massey Ratings.

Matt: [00:00:00] Hi, everybody. Welcome to discover college soccer today. We’ve got a special guest, a little outside the college coach realm, but one I know I’m interested in talking to, and that’s Ken Massey from Carson Newman university. Welcome. Or how does it say welcome coach? It’s I’m so used to it. Welcome Ken. 

Ken: I’m the furthest away from a coach.

You’re going to get, um, my little I have a 5 year old boy who’s learning, uh, just to kick the ball in the yard. That’s just about as far as I’ve gotten, but it’s an honor to be on your show, Matt. And, uh, I’ll probably learn something while we talk about the world of soccer and I’ll. Stick to the numbers 

Matt: well, and I appreciate you being here.

And that’s great. I think, uh, soccer journeys have to start somewhere. It’s usually in the yard. So I think you’re, you’re on the right track. Um, well, you’re, you’re a, um, a professor at Carson Newman. Your focus has been on. I’ll just say statistics, but, uh, I’ll let you give me a little bit of a, of a better primer on, on your background and how you got started into, to all this sports statistical modeling and what that looks like.

So, so please let, let us know about that. 

Ken: Yeah. Well, um, I did an undergraduate degree in math at A-N-A-I-A School Bluefield in Virginia. And then I did my graduate work at Virginia Tech, which is a D one school, and I’ve been at Carson Newman. Which is a D2 school for about 20 years now. And, of course, all sorts [00:01:30] of student athletes have come through my classes.

We’ve gotten to know some of them, soccer players, baseball, football, you know, we have, you know, probably 12 different sports on campus and I really appreciate. What the student athletes have to go through, I see how hard they work. They’re not going to get rich playing professional. Most of them. We’ve had a couple that went pro, but, um, for the most part, they do it for the love of the sport.

And whenever I can, I try to get out there and watch, um, like I said, I have some children now. We’d love to go watch the soccer games. It’s a great environment. And, uh, you can’t, uh, you can’t beat the atmosphere of just, uh, parents and coaches and friends just come out and watch people. They know, play, play their sport.

Yeah, absolutely. 

Matt: Well, you know, I have a soft spot for, uh, for, for math folks. My dad was a, was a high school math teacher for 30 years and taught some college statistics and stuff like that. So, so I definitely have a. An appreciation for, for your body of knowledge. So, so while at Bluefield and Virginia tech, which I was, I was this close to going to Virginia tech, I actually had my, uh, my deposit in, and I knew who I was rooming with, but last second change, uh, and I ended up going to Catholic U for my undergrad, but, um, so.

At, at Bluefield and Virginia Tech, what, what drew you to, you know, I think I can create a statistical model to, to rank sports teams. Uh, well, 

Ken: this was back in the mid to [00:03:00] late nineties when the Internet was, uh, just catching on and I had become interested in Virginia Tech football and. If you’ve ever seen the USA Today print newspaper for for decades, they printed the Sager in ratings and I was just curious to how that would work for college football.

I think that was the 95 season when Virginia Tech beat Texas in the sugar bowl. And I went to that game. And, um, I just. Piddled around with some of the ideas I had learned in math and stats class and said, you know, if I model teams as nodes in a graph and, uh, work out this, these systems of equations, uh, let’s just see what happens.

And so it was just really, um, exploratory just to see what I could come up with. And I ended up turning it into an honors project. And then a website. And it caught on with the for college football and over the years, I’ve just it’s just become my hobby to collect sports, uh, you know, soccer, basketball, football, tennis, anything you can imagine and just try to make sure that the model I have.

Works for every sport, every level, and, um, I’ve gotten a lot of feedback from the smaller schools, not the NFL’s and the NBA’s of the world, because they’ve got all their advanced statistics and they’re only 30 to 32 teams. But where these, uh, computer ratings really, [00:04:30] uh, provide value is when you’ve got hundreds of teams, like you would say, in college soccer.

Um, at all different levels, and they don’t play balance schedules. They’re very geographically dispersed. And you want to figure out relatively speaking, which programs are stronger than others for playoff seedings or for recruiting purposes. And so, um, I’m glad I started this a long time ago. I didn’t think I would do it my whole life basically, but, um.

It’s been a joy to me, and I’m glad I can have a website where people can see my work and provide feedback. I enjoy that kind of thing. 

Matt: Yeah. I mean, it is, it’s an awesome amount of information. Uh, and, and the crazy thing is, I mean, I could see it, you know, just starting from, we’ll say humble beginnings of college football ratings, but you go through your website right now and, and I can see, uh, you know, minor league baseball, uh, different soccer leagues, you know what I mean?

You got kind of everything under the sun, uh, on here. So how is it, again, now we’ll get, we’ll get a little bit geeky here, but. But how does that ratings methodology methodology change from sport to sport? Cause obviously something like soccer, uh, you know, where you’re playing games versus tennis, which is different matches.

And, and how does it change from sport to sport or what, what specifically you’re kind of looking at that, that what cuts across all sports, I guess. 

Ken: Yeah. Well, the, the [00:06:00] key is to understand the, an area of uncertainty in any sport. Soccer is probably going to have. More than most sports where, you know, you might get more shots on goal, but, you know, just like flipping a coin sometimes, which 1, which team gets more to go in.

Um, so the 1st step is to to look at a score and come up with a mathematical model of how certain you would be that the team that 1 is actually the better team I can soccer. If you went for 0. Uh, that’s pretty convincing. But if you in 2 to 1, the computer is going to realize, hey, it could have gone the other way very easily.

And, um, it collects every observation of every score accounts for home field throughout the season and it builds like this network. Or you can imagine, let’s say Carson Newman is attached to Tusculum and Mars Hill and all the teams they play and then they might be attached to teams further south. Let’s say into Georgia and those Georgia teams might be connected to teams in Florida.

And then, as you work your way around the country, even though any 1 team may only play 15 opponents, those 15 opponents played numerous other opponents and eventually they all get connected and they’re kind of pushing and pulling on each other. Uh, with this mathematical model, which tries to optimize, you know, the ratings that would kind of fit what actually got observed.

And, um, it with modern computers, it’s not too tough, but I try to, you know, my background is. Scientific computing, so I, I’ve taken [00:07:30] care to try to build my software. So that’s very robust to every sport that you could come up with. And it computes even I don’t do high school soccer, but I do high school football.

And there you’ve got over 15, 000 schools that get connected and it can run in just a couple of minutes. 

Matt: Wow. Wow. That’s awesome. Well, let’s talk, excuse me, specifically, uh, about the soccer ratings. You know, I’m, if I’m looking at, uh, you know, the D1 college women’s ratings on your site, you see, you see the rating and, and, you know, 1, 2, 3, whatever, but then there’s like a little number.

And I guess that’s the number that really is the, the key component here, right? Uh, that, that, that ratings number, but what, what do all these numbers and the different columns and, and all that kind of stuff really mean? Yeah. 

Ken: Uh, well, I have an overall ranking, which would be, you know, kind of to model, like, if there were an 8 people, and maybe there is for soccer.

I’m not sure. But if if there were an 8 people, where the humans actually paid a lot of attention, like, they do for college football, what would be kind of the way they would see it? Um, and then you’ve got a power rating, which is how good the team probably actually is, you know, like I said, some teams may have been luckier than others.

Um, and. The ranking might be like, might be the number 3 team in the country, but then their power rating is going to be how many goals per game better on average than that typical team. So those [00:09:00] power ratings are going to be. Real numbers. Whereas the rankings are just ordinal numbers. 4, 5. And then the power rating is subdivided into offense and defense.

Some teams, you know, uh, might win games four to three all the time. And other teams might prefer to keep it one to zero or two to one. So different styles would have different paces of play and that’s accounted for in the offense and the defense. And then I’ve got a couple of columns for strength of schedule, uh, for the season up to this point, and then for the, the full, uh, remaining schedule as well.

So, uh, it is a lot of numbers, but I think that that first ranking is pretty, pretty much like if you were to seed playoffs, what would be the fairest way to do it? Um, that first cop, that first ranking column you see. Okay. Well, 

Matt: in terms of. Um, you mentioned the AAP poll is a different thing and I don’t, they don’t think they really have a, an AP well, maybe they do for, for division one, but definitely not for two and three, although there’s, there’s a, and United soccer coaches does their regional rankings.

And then the big 1 that people talk about along with college football is RPI. Um, so, so how is like yours say different than RPI? Because I think that’s a computer, a computer model as well. Right, 

Ken: right. The RPI. Is a computer model, but it’s it’s limited. Um, remember a few minutes ago, I was talking about, like, [00:10:30] Carson and might play teams in North Carolina who play teams from Georgia play team from Florida.

The is only going to go 2 steps. They’re going to go opponents and opponents opponents, which limits how well it can see the whole country. So, my system. Effectively goes an infinite number of steps. It explores the entire graph. As far as it’ll go, um, and for the most part, teams are going to be connected at least within division.

Um, there, I wish there were more games between divisions say D1 versus D2, but at least my system tries the best given the data to separate those divisions. Whereas RPI wouldn’t be able to do that in that 

Matt: accident. That’s. A good point and question. I had one of the cool things that your site does or has the ability to do is, is matchups.

Um, and, and you can have, say, the number one D2 team. What, how is it going to fare against, say, the bottom D1 team? Because, because a lot of folks just automatically think D1. Must be better than D two must be better than D three, et cetera. But, but as you and I both know, there’s a lot of times when that’s not the case.

There’s going to be D two teams that can always beat D ones, et cetera. So does your. If I’m looking at the ratings for say, D1 and I see, you know, that a specific number there in that ratings column, you know, 8. 88 or something. And then the next, but, but the number 1 D1 team might have a rating of, let’s say, I [00:12:00] think Florida state here says 9.

64, but the number 1 D2 team might be a 7 or something like that. So, so is that, is your system taking into account that say difference in division? 

Ken: Yes, given the data that’s available, um, before we got on the call, just out of curiosity, I looked at division 1 went 26, 2 and 3 against the rest of the divisions.

So, they did dominate, but then. You can’t just even look at the schedule. You have to see what was it top tier D1 teams beating up on, you know, mid tier D2 teams, or was it bottom D1 teams beating up on top D2 teams. So the system is going to account for the strength of the teams within division that end up playing each other.

And it kind of moves like, if this is D1 and this is D2, whereas RPI couldn’t really tell which one’s better. My system by looking at all those connections is going to kind of. Let them balance like this. So it’s, you’re going to get an overlap, like, like if this is D one and this is D two, I don’t know, I’d have to look, but there’s going to be an overlap where maybe the top D one team might be ranked 140.

And I mean, I said it backwards. The top D two team might be ranked say in 140 and D one, right. And those numbers are directly comparable. 

Matt: Okay. Yeah, that, that makes a lot of sense. I mean, just obviously there’s a crazy amount of data and numbers being [00:13:30] crunched on the background, but. I mean, I have a hard enough time trying to You know, keep up with, with my daughter’s high school schedule and stats that come out of her games.

So, I mean, how are you keeping track of all this stuff and all the changes? Cause I mean, there’s probably a handful of schools that are switching from D2 to D1, uh, over the summer. And, and so you gotta make those adjustments and, and, and that’s just soccer. I can’t imagine all the other sports and all that amount of data you’re trying to, to keep up to date.

How do you do that? 

Ken: Uh, well, the 1st thing is to keep my sanity. I don’t worry about statistics. I don’t keep track of goals, shots on goal or, you know, and other sports time of possession or 3 point percentage. Any of that other stuff. It’s just the score of the game where the game was played and, um. I have different websites, like the has some information and then the junior college sites that I can get information from each year.

I try to set up the schedule as best as I can and then, uh, weekly try to kind of go through and make sure that. If people have emailed me corrections, like this team switch conferences or something like that, I’ll try to fix them manually. Um, and then a big annoyance is schools that have similar names.

Like, you might have. 3 Saint Francis’s, or something like that, and they’re always getting mixed up. Uh, so that that part of the job, I don’t really enjoy, [00:15:00] especially when you’ve got teams and different divisions that have similar names. Um, but I try to keep it as clean as possible. I would say 80 percent of my work is data.

Data gathering and organization and cleaning and things like that. But I’ve over the years, written software tools that helped me do that fairly 

Matt: efficiently. Okay. I said, there’s gotta be some sort of automation, uh, in there, but I can imagine the surfing through all the different trinities, St. Francis and St.

Mary’s of the world. It’s gotta be a tough, tough road. Oh, well, and I mean, I. This is, this will be come a little bit differently, but normally on these, I ask a final question to my college coaches for, for advice around the college recruiting process. And I’m going to ask you the same question, but a little different bent because you’re, you’re a college professor, you get to be teaching these kids and see them on a day to day basis.

So, you know, as a professor, as a sports fan, you know, what advice do you have for any high school students right now that are, that are going through their college search process? And it doesn’t have to be sports related at all, just some good. Good, hard, honest advice from a, from a college professor who’s seen successful students and unsuccessful students.

What advice would you have for anybody? Um, 

Ken: well, and this would apply academics as well as sports is try to find the right level. You don’t want to be. I mean, it sounds weird. You don’t want to be the [00:16:30] best player on the field because you’re not you’re not going to be surrounded by like teammates and opponents that would challenge you.

Um, now, somebody’s got to be the best, obviously, but go to a place where you feel like you can can grow and be challenged. Um, and that doesn’t have to be, uh, based on any kind of, um, I guess, human. Categorization like D1 versus D2, uh, talk to the coaches, look at the numbers to see, uh, you know, are they getting recruits that I can go in and compete with and learn from, um, and also pay close attention to geography.

I’ve noticed that Carson Newman just, uh. A lot more spirited games and competition when there are a lot of local kids. So if you’ve got a local school, that is D 2 or D 3, and they’ve got a good reputation there. I got a good coach. Uh, don’t, uh, don’t ignore them and try to go out of state. Just so you can say, you play D1, uh, geography.

I think is important as long as you’re being challenged on the field and in the classroom. Um, so make sure you can find a major program that you want to, uh, stay and graduate with. 

Matt: No, absolutely. Couldn’t agree more. Well, Ken, I really appreciate it. Want to say thank you for the work [00:18:00] that you do and keep it up because I know I reference your site quite often and have it linked off the discovercollegesoccer.

com website for folks to see because that’s a question I get is, you know, how do you tell, especially in something like division three where there’s just hundreds of schools? How can you really tell? And this is a great question. A great barometer, at least for the on the field stuff off the field stuff.

You got to go somewhere else, but, uh, we appreciate what you do and really thank you for your time today. Thank you, Matt. Take care.