# Indexing

Indexing is a common extraction method used in hunt puzzles. It involves using particular numbers to select the Nth letter of a given string of letters. It is likely the most common extraction method across all puzzle hunts, due to its versatility and accessibility, especially to new solvers.

## Puzzle Application

At a basic level, indexing is very simple. Numbers can be paired with words or phrases in order to extract individual letters from each. For example, if a certain index is 6, then one would count 6 letters into the word/phrase/string and write down that 6th letter.

```Shopping List: 2 Bananas, 12 Hamburger Buns, 5 Radishes, 1 Watermelon, 4 Live Chickens, 3 Parsnips
BANANAS (2), HAMBURGER BUNS (12), RADISHES (5), WATERMELON (1), LIVE CHICKENS (4), PARSNIPS (3) --> ANSWER
```

These numbers can be pre-paired with the strings as seen above, or the numbers themselves may be determined through additional steps, further obscuring the indexing revelation. One way to do this is by having answers associated with particular numbers, and being indexed into by their appropriate number.

```SNIVY TYPHLOSION RILLABOOM IVYSAUR FROGADIER
```

In this case, each Pokémon gets indexed by its local Pokédex number: SNIVY (1), TYPHLOSION (6), RILLABOOM (3), IVYSAUR (2), FROGADIER (8) --> SOLVE.

Extremely simple types of indexing are considered obvious enough to not need numerical clues (or clues at all.) These simple examples include initialization (index of 1), diagonalization (increasing index of 1, 2, 3, etc.), centralization (indices which always point to the middle letter of a word with odd length), and terminalization (indices which always equal the word length, pointing to the last letter). Centralization and terminalization are usually given verbal clues, while initialization and diagonalization are most often unclued.

Indices can be presented at the outset of the puzzle and paired up with other words through the course of solving. A common structure for index-pairing involves the solver being given two sets of clues, with one set being ordered and the other set being given indices. In that situation, the clues must be paired so the indices can be correctly positioned, and the solver is usually expected to apply the indices to the half of the clues which started unindexed.

Some type of indexing always forms the extraction step in identify, sort, index, solve puzzles.

## Strategy

The action of indexing requires very little strategic thought. Two elements of indexing, however, have the tendency to make problems for solvers.

### Is It An Index?

Whether or not a particular number or set of numbers found in a puzzle is actually meant to be used as an index can be a difficult question to answer, particularly if there are a lot of different sets of numbers being thrown around.

One common confusion can be whether numbers presented after clues are indexes or enumerations. An easy way to differentiate them is to remember that indexes will usually be single numbers. If any of the clues is followed by two numbers separated by a space, it's likely that it's an enumeration for a two-word phrase, and the rest should be treated as enumerations as well. Alternatively, if the answers to the clues consistently end up being the same lengths as the numbers, it's likely they're enumerations rather than an overly-signposted terminalization extraction. If the numbers span from 1 to some number uniquely in some order, they might be a method to reorder the clues or elements instead of an index.

Another possible confusion is confusing indices for other extractions involving numbers. If numbers don't work as indices, then they could be used as caesar shifts or A1Z26.

When presented with multiple numbers for a given set of information, it can be difficult to know which number is an index, if any. Since indexes can't go beyond the length of their given string, this can sometimes be used to narrow down possible index number sets. If any of the numbers in a set can't be used as an index on its respective string, then it's unlikely any of the others will be used either. It is possible that the string is incorrect, so make sure to double-check your answers and your numbers to make sure you're not excluding anything that shouldn't be excluded.

As an example for when numbers for indexing can be ambiguous, consider the list of Pokémon above:

```SNIVY TYPHLOSION RILLABOOM IVYSAUR FROGADIER
```

One can rule out, for example, the National Pokédex numbers or base stat totals for those Pokémon because they are extremely large compared to their name length. However, since the intended indices for this example-- regional Pokédex numbers-- are obscure to solvers unfamiliar with Pokémon, it may take a very long time for those solvers to identify the proper indices at all (or even to identify this as a puzzle requiring indexing.)

### What Gets Indexed?

When a puzzle provides a lot of streams of information and expects solvers to index into a specific subset of that information, it's very easy to get confused. A good example of this is music identification puzzles, in which each track has an album, artist, title, and track number that each have the possibility of being relevant.

One of the strategies to deal with this problem is the same as one used for determining which set of numbers to use as indices. If a set of known indices contains numbers too large to index into one of your datasets, then it's likely that it's not meant to be indexed into. As before, it's still possible either the indices or the strings are incorrect, but that's what double-checking is for.

Since this kind of situation is unlikely to occur without a way to reliably determine the correct path, taking note of the different streams of information and the extent to which they've been used during the rest of the puzzle is a good idea. Information is usually used somewhat equally throughout a puzzle, and if a piece of information hasn't been used at the time of indexing, it's likely that that information will be used either during indexing or for an unknown step beyond indexing. Either way, it'd be wise to confirm which is the case.

## Notable Examples

### Played Straight

• Quick Fix (Silph 2021) (web) - Requires pairing up clues Click to revealwith the same solution. One clue contains the number used to index into the answer, and the other clue is used for ordering.
• Over 9,000 (MITMH 2021) (web) - A straightforward indexing puzzle, just with rather large index numbers Click to revealand reuse of the same indexes multiple times on different strings.

### Notable Twists

• Interpol (BANG 23) (web) - Uses fractional indexes, requiring solvers to Click to revealtake a letter "between" two other letters. In this case, that meant counting the distance between two letters and taking the noted letter in that sequence. For example, if an index was [2 4/11], and the string had a second letter of P and a third of E, the index would be of the 4th letter of 11 between P and E in the alphabet, or L.
• Just Index (MITMH 2021) (web) - Uses overlarge, negative, and complex indexes. The trick ends up being that Click to revealthe strings can be found amongst the clues, Wordsearch-style. This means that indexes can go beyond the ends of entries, before the beginning of them, and travelling in directions perpendicular to the entries to accommodate the too-long, negative, and complex indexes, respectively.