In the following you can find more explanations regarding the outputs. For even more details (e.g. how the input files work or how to manually add more data) see the complete repository including the code and data (note that currently some of the explanations are only available in german).

Also you find a way of asking questions, share ideas and note bugs/errors.

Normal outputs #

In front of each table you’ll find precisely what constraint was added here. The mentioned episode always referrs to when the outcome of the constraint was revealed. In case of matching ceremonies there is also a note before the match how often they sat toegther in the past.

Everything related to I (Information content) / H (Entropy, after the note how many possibilities are still left) is the attempt to get a feeling for how much this decision was worth and how far they are from the finish line. This is related to information theory.

I[l/bits]: Represents how much information was gained with this decision assuming the given amount of beams are lit. Via \( 2^{-I} \) you can calculate back to the probabilities if you can relate more with that.

E[l]/bits: Is the expected value of the gain of information

Regular tables #

The fontcolor is an indicator for how high the probability for this match is (below 1% red, higher than 45% yello, higher than 55% cyan and more than 80% green).

The backgroundcolor shows which people have the highest probability with which other people.

• green background: Match is for both persons the most likely one.
• red/light gray backgroung: Match is for the person whose column/row this is the most likely one.

Summary table in the end #

In the end there is a summary of all constraints. A star in this tale means this match was formed the first time here. A small overview over the columns which might not be clear on first sight:

• L the amount of lights
• I refer to above
• new counts how many matches never sat together a matching ceremony
• min dist the distance is denoted with the amount of different matches, this column shows which other matching ceremony is most alike this one (and how much alike they are). This cannot be calculated for the first ceremony obviously.

Tree #

In the tree the first row (correlated with the person out of setA) at a level is always fixd. Thus, each level represents the matching of one (or multiple) people of setB to this fixed person of setA.

Already known matches (be it from truth booth or throug elimination) are put on the topmost levels. For the remaining levels the ordering is sorted based on the amount of different matches for the fixed person.