Ladder System

Prior to 2007, TUC leagues all used round-robin scheduling, with the league typically split into 6-team tiers when there were too many teams for round-robin to work well. Many of us experienced problems with these tiered leagues, including:

  • teams being "trapped" in a tier where the level of play was too high or too low for them, resulting in a series of lopsided games
  • delays caused by trying to schedule five weeks of games at once
  • inflexibility in rescheduling games (rainouts and cancellations), since results must be in by the end of the round

The ratings ladder was developed by the original Leaguerunner writers at OCUA in response to these and other problems, and retained in Zuluru. TUC first experimented with the ladder in the Monday Premier Co-Ed Indoor league in the winter of 2007. It was deemed successful enough to try on Monday and Tuesday nights during the summer of 2007 and in several of our fall offerings, and has since been adopted for virtually all TUC leagues.

The stated goals of the original developers were to:

  • prevent blowout wins/losses by matching teams of roughly equivalent ability
  • ensure that teams see a variety of opponents throughout the season
  • ensure that the "best" teams rise to the top of the division so that an overall winner can be determined

How It Works

Division of leagues into multiple tiers is no longer done. Instead, all teams are placed on one "ladder". Each team is assigned an initial rating, based on last year's performance, or some other criteria for teams with no history. For example, teams that finish in the top four one year might be given a rating of 1500 the next year, the next four teams rated at 1450, and so on. From this point on, the league standings are based on the ratings of the teams.

Each week, the schedule is created by matching the top rated team against the second highest rated team that they have not played recently. ("Recently" normally means about four weeks, but can change depending on the league specifics.) Then, the next team down the ladder that has not yet been scheduled is assigned an opponent in the same way, and so on, until all teams have an opponent. (This is a bit of an over-simplification; there is a small bit of randomness introduced into the process, and many possible schedules generated, with the chosen schedule being the one that minimizes the seeding differences between opponents across the entirety of the division.)

When game results are reported, a formula (see below) is used to determine the number of rating points which are transferred from the loser to the winner. The number of points transferred is based on the ratings of the two teams involved and the final score; beating a team rated above you is worth more than beating a team rated below you, and winning by a larger margin is worth more. Note that it is quite common that teams with a "better record" will be ranked below a team with a "worse record"; it's not so much whether you win, but who you beat.

Our administrators and conveners typically try to schedule games in rating ladder leagues about a week and a half in advance. This gives you extra time to know when and where your game is, but means that last week's results were not taken into account when this week's schedule was made.

Why It's Better

The single ladder makes it possible for teams to move up or down very quickly, which allows teams to reach their appropriate level in just a couple of games. This is an improvement over tiered systems, where new teams or teams that have had significant turnover since last year have been known to play their first round (five games) or even two rounds (ten games) against teams that are at a vastly different skill level. This would happen most commonly in lower tiers.

On a similar note, there were times where a particular tier was very competitive, but the "five week shuffle" demanded that two teams move up and two move down. When a 2-3 team moved down or a 3-2 team moved up, they were often moving into a tier where they did not have such close games as they had enjoyed. With the ladder, teams that are very competitive with each other should remain close to each other in the ladder, and end up playing each other a few times during the season. These close, competitive games and rivalries are what we all love, and what the ladder system tries to encourage.

By scheduling one week at a time instead of five weeks all at once, and eliminating all of the communications involved with organizing the shuffles, the work load for administrators and conveners is lessened and evened out. Delays at the beginning of each new round, where game locations for the first game were sometimes not known until the day of the game, are a thing of the past.

We've all had situations where injuries and vacation all seem to hit at once, and we simply cannot get enough people out to play a game. In the tiered system, if the game is to be rescheduled, it had to happen before the end of the round. In a ladder system, that game can be replayed any time (although, earlier is better, so that the results can be taken into account in future scheduling), or it can even be skipped entirely with minimal effect (in very rainy summers with multiple rain-outs, some teams have ended up playing only 14 games instead of 15).

The "Wager Ladder" Formula

(Copied from the OCUA web site)

The current formula is called the "Wager Ladder". In this system, games are worth a variable total of points, based on the final score. Each team contributes some number of points from their ratings value towards this total, but the amount they contribute is not fixed -- it depends on their relative ratings. This contribution is that team's wager contributed to the pot. The losing team wins back the number of points they scored, and the winning team takes the remainder of the pot.

Here's how it works in detail:

  1. A team's percentage chance to win is computed based on the pre-game ratings values. This will determine what percentage of the game's value the team must contribute. For example, if both teams are evenly matched, they will each contribute 50% of the total. If instead, Team A has a 60% chance of winning, they contribute 60% of the total value.
  2. After the game is complete, the pot value is computed. The pot value is double the winning score plus 10. For a normal game to 15, this would make the pot worth 40 points. For a game that ends in a timecapped win at 12, the pot would be 34 points. A game ending 17-16 after an overtime universe point would be worth 44 points.
  3. Each team's contribution is calculated by multiplying the pot value by the team's percentage chance to win. This represents the maximum number of ratings points that team could lose.
  4. The losing team then gains back the number of ratings points equal to their score. In an evenly-matched game ending 15-10, the losing team would gain back 10 points for a total loss of 10.
  5. The winning team then gains the remainder of the pot. So, for example, if their contribution was 20 points, and they won 15-10, they would get a pot of 30, for a net gain of 10.