Does "home field advantage" exist? Should we account for it?
No, we should not consider where games get played -- every game has the same importance and every team performs the same regardless of where the contest takes place.
This is a simple philosophy, and of course works great for intramural leagues, near-empty or split stadiums, and so on. It's also the philosphy used when we build division and conference standings -- no bonus points for winning on the road, no penalties for losing at home, etc.
There are several computer rankings that ignore home/away/neutral sites: Colley, Wilson, Rothman, ..., to name a few. Every game is treated as if it was played at a "neutral site."
Yes, home "field" provides an advantage for one team over the other, and so maybe we should account for where a game gets played. One simple way to do this is to have a generic home advantage bonus for an entire league -- for example, add 3 points to every home team in college football, 2-3 points for college basketball, 0.25 goals for ice hockey, and so on.
Others, like Ken Pomeroy (kenpom) or Kenneth Massey, calculate a separate home advantage for every team in a league. As of 27 December 2023, KenPom says West Virginia has the best men's basketball home court advantage (4.6 points), and Navy the least (1.1 points). Massey's have a smaller range, from 2.7 to 3.1. But within a point or so, they all hover around 3 for men's basketball. For this reason, I go with the "generic advantage" philosophy, rather than create one for every team, but that's my personal choice.
The NCAA does this in their RPI calculations in a very simple, clean way: since they don't care about points scored or allowed, they account for location in the win/loss columns: a road win actually counts for 1.4 wins(!), and a home win counts for only 0.6. So a team that has gone 3-1 in four home games would actually only be 1.8-1.4, according to them.
• Want to create your own ranking?
• Want to browse other ranking systems?
• ...including this author's?
• More to add here...