10U Rankings
Ranked from game results: who beat whom, and how good they were. Ranked by strength rating, and no team more than 2 to 5 spots above the best team it has beaten (more spots for more wins near that team). Click a team for its page. Updated Nov 17, 2026.
11U Rankings
Ranked from game results: who beat whom, and how good they were. Ranked by strength rating, and no team more than 2 to 5 spots above the best team it has beaten (more spots for more wins near that team). Click a team for its page. Updated Nov 17, 2026.
12U Rankings
Ranked from game results: who beat whom, and how good they were. Ranked by strength rating, and no team more than 2 to 5 spots above the best team it has beaten (more spots for more wins near that team). Click a team for its page. Updated Nov 17, 2026.
13U Rankings
Ranked from game results: who beat whom, and how good they were. Ranked by strength rating, and no team more than 2 to 5 spots above the best team it has beaten (more spots for more wins near that team). Click a team for its page. Updated Nov 17, 2026.
How the rankings work
Battle Rankings rank Oregon travel baseball teams from their game results, one age group at a time.
How the method works: the Bradley-Terry model
The rankings use the Bradley-Terry model, a standard method for ranking from head-to-head results. It is named after Ralph Bradley and Milton Terry, who published it in 1952, and many chess and sports rating systems use the same idea.
Every team gets a rating, a strength number. When two teams play, the chance the first one wins is
rating of team A ÷ (rating of A + rating of B)
So a team rated 4 playing a team rated 1 is expected to win 4 out of 5 games. An average team is rated about 1.
The ratings are worked out all together, so that they explain every result of the season as well as possible. What that means for one team: add up its chance of winning each game it actually played. That is how many games a team this strong would be expected to win against this exact schedule. The ratings are set so that this number equals the games it really won (a tie counts as half). If a team had won one more game, its rating would have to go up until the two matched again. Every team's “Why” tab shows this check, game by game.
- Who you beat matters. A win over a strong team was not expected, so it lifts the rating a lot. A win over a weak team was expected, so it lifts it a little.
- Who you lose to matters. Losing to a strong team costs little; losing to a weak team costs a lot.
- Teams that mostly play each other. Their wins and losses against each other mostly cancel out. Their games against teams from outside the group are what move them.
- One game is one game. A team that beat you once can still be ranked below you if the rest of the season says it is weaker. The rankings weigh every game.
Our changes to the method
We make two changes to plain Bradley-Terry.
1. One pretend tie for every team (statisticians call this a prior, or “shrinkage”). Every team starts with one pretend tie against an average team, and it counts in the check above. Without it, a team that went 5-0 would get an endless rating, however weak its opponents. With it, every rating is pulled a little toward average: a short perfect record still counts, but a team has to keep winning to climb. It also keeps teams that mostly play each other near average until they beat teams from outside the group.
2. Win your way up: 2 to 5 spots above your best win. Sorting teams by rating gives the rating order. Then one rule is applied: a team can be ranked only a few spots above the best team it has beaten. Only outright wins count, not ties. A great record against weaker teams earns a high rating, but to be ranked near the top a team has to go beat teams near the top.
- Start at 2 spots. One win over a team lets you be ranked up to 2 spots above it.
- Earn more spots by winning more. Each extra win against your best-win team or the 4 teams ranked just below it adds 1 spot, up to 5. Beating the same team counts at most twice.
- Example: a team went 7-0, so its rating says #3. But the best team it beat is ranked #8, and it has no other wins near #8. It can be ranked no higher than #6 (8 minus 2), so it is moved down to #6.
- Why it grows: the top teams should not be able to stay on top by refusing to play. A team ranked #6 that beats #5, #7, #8 and #9 has earned 5 spots above #5, so it can climb all the way to #1 without the top 4 ever agreeing to play it.
- No wins yet: a team that has not beaten any ranked team is not limited by this rule. Its rating alone sets its place.
- Settling: when a team is moved down, the teams it passes move up one spot each. Every team’s spots are then checked again against the new order, and this repeats until the order stops changing.
We tested these choices on the 2026 season by hiding games, ranking without them, and checking who won. The higher-ranked team won about 72% of the games the rankings had not seen. The win-your-way-up rule costs about 1 game in 1,000, and we keep it because a spot near the top should be won on the field. About 1 game in 4 is an upset no matter how teams are ranked: that is baseball.
In one line: Bradley-Terry with a one-game prior, plus a best-win rule that starts at 2 spots and grows to 5 as a team keeps winning, using wins and losses only.
Which games count
- Only who won. Each game counts as a win, a loss or a tie. The score does not matter: a 1-run win counts the same as a 15-run win, so nobody gains anything by running up the score.
- Only games inside the age group. A 10U team is ranked only on its games against other 10U teams. Games against older or younger teams show on the team's page but do not count.
- One season at a time. Each season runs January 1 to December 31 and starts fresh. This season covers the spring and summer travel season; fall ball is not included.
- At least 5 games. A team needs 5 games against teams in its age group before it is ranked. Teams with fewer still have a page, marked "not ranked yet".
- Only games against ranked teams. Games against teams that are not ranked yet don't count either: too little is known about those teams to say what beating them means. They still show on the team's page.
- Where the scores come from. Final scores are collected from GameChanger and checked against both teams' schedules.
Why not just count wins?
Say Team A beats Team B, Team B beats Team C, and Team C beats Team A. Which team is the best? Each one has a win and a loss, and each one beat a team that beat the team that beat them. Simple rules can't settle it: win-loss records treat all three the same, and "you're ahead if you beat them" goes around in a circle. Real seasons are full of these circles, and teams play very different schedules: one team goes 15-2 against weaker clubs while another goes 9-8 against the best in the state. Most teams never play each other at all.
So the rankings look at every game at once. Every result says something about both teams, and wins over strong teams say more than wins over weak ones. The Bradley-Terry model above finds the strength numbers that agree with all the results as well as possible. In a perfect A-B-C circle, the teams would come out even, and the rest of their games decide who is ahead.