Back to Records Index Up to Feedback Index Chris' Home Page

Feedback - Football:All-Time Win/Loss Records


The web site rolltide.com, list games for the University of Alabama back to 1892. Your site does not include these. Is there a reason?

Fred Cook (12/04)

Response from the editor:

There are actually two different databases underlying the queries on my web site. One is my own, which contains only annual won-lost records by team. I do include Alabama's records back to 1892, and you can see them here:

http://football.stassen.com/cgi-bin/records/fetch-team.pl?team=Alabama#1892

The other database which my query forms use as a source is James Howell's data. His is a much more detailed database that has all-time game results -- score, date, location, etc. Every form on my web site which uses James Howell's data includes this disclaimer:

Note: James' site contains only data for years when each team was considered 'major' (equivalent to today's I-A), and therefore will not necessarily yield true 'all-time' results.

You can see James Howell's Alabama data here:

http://www.jhowell.net/cf/scores/Alabama.htm

James only counts Alabama as "major" starting in 1902. I do not know how James made those determinations, and since he is courteously providing his data to me, I am not really in a position to argue with him about it.

If you want a query similar to mine on James Howell's data, but you really need true all-time results, I would recommend you instead use the College Football Data Warehouse web site. Their Alabama page is here:

http://www.cfbdatawarehouse.com/data/div_ia/sec/alabama/index.php

I helped improve some of CFBDW's database queries, writing a "vs-conference" query that uses the conference membership at the time the game was played, and adding start-end restrictions on various all-time queries. David DeLassus is a nice guy, and his database is quite useful. In fact, I have been thinking about taking down my own database and queries, and just leaving a pointer to CFBDW.

Through past research I have that the Univ. of Michigan is the all-time winningest team in NCAA college football history period please, please respond your site only went back 100 years.
Thank You

Kwadwo Joyner (6/98)

Response from the editor:

"All Time Winningest" means "has the most wins." Michigan does have more wins than any other I-A college football team. The team with the most wins may not be the one with the highest winning percentage, though.

At the time of this writing, Michigan has 23 more wins than Notre Dame... but since the Wolverines have played 44 more games (and have six less ties), they also have 27 more losses than Notre Dame. The result is that Michigan's all-time winning percentage is currently (1998) about 1.25% lower than Notre Dame's.

Incidentally, this site has all-time data. If you go to the "compute request" form and just leave the default values in, you will get all-time winning percentage. The very first college football games were played in 1869.

In reviewing your season by season records for some teams that I was interested in (by the way, great job with this site), I noticed what I think is an error: 1986 Virginia Tech you have listed as 9-2-1. I believe that the official record has been revised to 10-1-1, due to a game forfieted to the Hokies by Temple University for using ineligible players. I don't know what your standards are for the presentation of this data, i.e. do you count forfiets? If so, you may want to take a look at Temple's record for that year as well.

Rusty (3/98)

Response from the editor:

Thanks for the tip! I checked with Temple; an assistant SID informed me that all six of Temple's wins were forfeited that year. The database has been adjusted.

Whether or not to adjust for forfeits is a matter of some debate. If forfeits are ignored, non-compliance with the rules is essentially rewarded. If forfeits are counted, the database no longer reflects actual on-the-field results. I believe that the latter is the lesser of two evils, so forfeits are counted in the win percentage calculations. I show both the original and adjusted record in the team's database, however.

Each team's year-by-year records can be viewed from here. For example, you can see the change to Virginia Tech's record here. The line in red gives the record before adjustment for forfeit. Currently these lines are ignored when computing winning percentage.

The w/l record program is cool. My only suggestion would be to somehow factor in some kind of strength-of-schedule comparison and see how the numbers come out. It would take a lot more number crunching of course. Maybe your idea about the NCAA would work better for this. Thanks.

Sid (10/97)

Response from the editor:

You make a very good point. A ranking by winning percentage is only meaningful if the teams can be assumed to face similar levels of competition. For that reason, it is important to be careful about what data is counted and what is not. As a very rough approximation, I use only results earned in the highest division of play. Play in lower divisions is entered into the teams' database, however.

Each team's year-by-year records can be viewed from here. For example, look at the recent year-by-year records of Yale. The lines in blue represent play in lower divisions. Currently those lines are ignored when computing winning percentage.

If we were to attempt to fully adjust winning percentage for strength of schedule -- even for differences between various I-A teams' schedule strength -- we would essentially be inventing a power ratings system (and ranking teams by that power rating). I do not have game-by-game score data online, and I think it would be very difficult to derive sensible ratings for the early years of the sport.

I just found your site today, and think its outstanding. One problem however is I don't think Marshall is a division I-A football school, but they show up when you list the top teams by winning percentage. I think they are I-AA and shouldn't be included in the list for I-A schools.

Anonymous (9/97)

Response from the editor:

Marshall just re-joined I-A effective this season (1997), as a member of the MAC. Also, Marshall was a member of the MAC in the 1970s and earlier. In between Marshall was, as you note, in I-AA. I have fixed the database so that their I-AA games aren't counted towards winning percentage. See the notes area (I-AA and other) for a discussion of database sanity versus division-hopping teams.

Do you have plans to do the same thing for NCAA basketball?

Stan Hiltman (6/97)

Response from the editor:

I don't have current plans to expand into other sports, for three reasons:

  1. The college football database is not really complete.
  2. I don't have much data for other sports lying around.
  3. I really don't have the time these days.
I hope that the NCAA sees this site, decides that it is a neat idea, and eventually puts their statistical library online (for all sports).

Back to Records Index Up to Feedback Index Chris' Home Page