Advanced Topics¶
MIT-Tab’s TabSetting module allows tournaments to modify the internal logic of how mit-tab works. This article will document ways that you can use tab settings to improve your tournament.
For each of the modifications below, you will have to know how to create or modify a tab setting. In order to do that, first, navigate to the admin interface in the navigation menu (”Admin” > “Admin Interface”)
You will then be taken to a page titled “Site administration”. On that page, click the “Tab settings” link. You should now be on a page with the heading “Select tab settings to change” that looks like this:
Modifying a Tab Setting¶
To edit a tab setting, click on the one that you want to modify and change the “value” attribute
Creating a Tab Setting¶
To create a tab setting, click the “Add tab settings” button in the top right corner of the page. The “key” should be the (case-sensitive) name of the setting, and the “value” should be the desired value for the setting
Default Tab Settings¶
Here are the default TabSettings unrelated to the power-pairing calculations
(those related to that can be seen at the bottom of this page.)
Note: Some may already exist on your server, some may not. The code will use these settings if provided regardless of if it showed up in the interface before you created it
| Name | Default value | Purpose |
|---|---|---|
| cur_round | 1 | Control the current round of the tournament. Don’t modify this |
| tot_rounds | 5 | Number of in-rounds at the tournament |
| min_speak | 0 | The minimum speaker score allowed to be submitted by the tab staff |
| max_speak | 50 | The maximum speaker score allowed to be submitted by the tab staff |
| max_eballot_speak | 35 | The maximum speaker score allowed to be submitted by an e-ballot Note: This value will be allowed. In other words, if the value is 35, a 35 does not have to be justified to tab staff but a 36 does |
| min_eballot_speak | 15 | The minimum speaker score allowed to be submitted by an e-ballot Note: This value will be allowed. In other words, if the value is 15, a 15 does not have to be justified to tab staff but a 14 does |
| lenient_late | 0 | The latest round where people who do not show up are given average speaks instead of speaks of 0 by default. Note: This only applies _before_ the round. If you change this after the round was paired, it will not take effect. You can manually allow a lenient_late in the admin interface. Docs coming soon. |
| pairing_released | 0 | 1 if the pairings are publicly visible, 0 when they are not |
| fair_bye | 1 | 1 if only unseeded teams should be eligible for a first round bye, 0 if all teams should be eligible |
| use_team_codes | 0 | 1 if team codes should be display on the public pairings view, 0 if normal team names should be display |
| teams_public | 0 | 1 if the teams list is publicly visible, 0 if not. |
| judges_public | 0 | 1 if the judges list is publicly visible, 0 if not. |
| debaters_public | 1 | 1 if the debaters are publicly visible on the public teams list. |
| team_codes_backend | 0 | 1 if the team codes are used for all non-public views, 0 if not. |
| nov_teams_to_break | 4 | The number of novice teams to be included in the break (exclusive of novice teams who are varsity breaking) |
| var_teams_to_break | 8 | The number of varsity teams to be included in the break. |
| nov_panel_size | 3 | The number of judges on a novice outs panel. |
| var_panel_size | 3 | The number of judges on a varsity outs panel. |
| var_to_nov | 2 | The offset of the novice break to the varsity break. If novice semis happen when varsity quarters happen, the offset should be 1. If novice semis happen when varsity octofinals happen, the offset should be 2. |
| var_teams_visible | 256 | The number of teams above which the varsity outround is visible. For example, if it were 8, quarterfinals and above would be visible, if it were 4, semifinals and above would be visible. |
| nov_teams_visible | 256 | The number of teams above which the novice outround is visible. For example, if it were 8, quarterfinals and above would be visible, if it were 4, semifinals and above would be visible. |
| gov_opp_display | 0 | A toggle to switch the outrounds view from displaying Gov and Opp (when value is 1) and Team 1 and Team 2 (when value is 2) |
| sidelock | 0 | A toggle to indicate whether you are side-locking outrounds. If 1, then the gov opps for outrounds will not be random, but will adhere to sidelocks, and will be indicated on the pairing card, and bolded on the front-facing pairing display. |
| choice | 0 | A toggle to indicate whether you would like to display who has choice, whether you would like this to be be indicated on the pairing card, and bolded on the front-facing pairing display. |
Running a Tournament with a Non-Standard number of rounds¶
To run a tournament that has more/less than 5 round, change the value of the
tot_rounds setting to the number of rounds. If you don’t do this, the
gov/opp count optimization will not work, and pairings may get messed up.
Dropping a Team with Average Speaks¶
By default, when teams forfeit, they are given speaks of 0 and ranks of 7.
To override this behavior, create a tab_setting with the key lenient_late.
The value of the tab setting is the last round number where the default
behavior of a forfeit is average speaks. For example, setting it to 1 would
default any forfeits in round 1 to use average speaks. Setting it to 4 would
make it the default behavior for rounds 1-4.
To override the default behavior for the round, navigate to the admin
interface and click on the “No shows” link. You will now see a list of links
formatted as {Team name} was a no-show for round {round number}. Find the
team and round that you want to override, check/uncheck the “lenient late”
option, and click “save”.
Modifying the Pairing Algorithm¶
The pairing algorithm uses creates minimum weight maximal matching algorithm to create the round. You can read more about the algorithm used here. What this means is that the pairing algorithm isn’t a list of specific rules. Instead, it generates every possible pairing for the round, gives a score to each pairing, and then selected the pairings with the highest score. Penalties with the highest magnitude are the “worst” penalties. Those with values closer to 0 are less significant.
Note: The value of the setting must be negative. Otherwise your pairings will be very bad.
Here are the settings you can create, along with their default values:
| Tab setting name | Default value | Description |
|---|---|---|
| power_pairing_multiple | -1 | For each position that the pairing is shifted from the “ideal” power-pairing, it’s multiplied by this number |
| high_opp_penalty | 0 | One team will have to opp more than half of the rounds + 1 (e.g. 4 opps at a 5 round tournament) |
| high_gov_penalty | -100 | One team will have to gov more than half of the rounds + 1 (e.g. 4 govs at a 5 round tournament) |
| higher_opp_penalty | -10 | One team will have to opp more than half of the rounds + 2 (e.g. 5 opps at a 5 round tournament) |
| same_school_penalty | -1000 | Both teams are taking protection from the same school |
| hit_pull_up_before | -10000 | A team who has hit the pull-up already will hit the pull-up again. (Note: Teams that have already been pulled up will never be considered for the pull-up) |
| hit_team_before | -100000 | The two teams in the pairing have already debated each other |
So, if I wanted to change the penalties so that having 4 govs was half as
important as it usually is, I’d change the value of high_gov_penalty to
-50. If I wanted to make it not considered at all, I’d change it to 0.
If you would like to prevent seeded teams from getting the bye, set
fair_bye to 1