formats.Rmd

---
title: "Formats of Competition Results"
author: "Evgeni Chasnovski"
date: "`r Sys.Date()`"
output: rmarkdown::html_vignette
vignette: >
  %\VignetteIndexEntry{Formats of Competition Results}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
---

```{r setup, include = FALSE}
knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)
```

`comperes` offers a pipe (`%>%`) friendly set of tools for storing and managing competition results (hereafter - results). This vignette discusses following topics:

- Storage of results.
- Conversion between different formats of results.
- Notes on package application.

Understanding of __competition__ is quite general: it is a set of __games__ (abstract event) in which __players__ (abstract entity) gain some abstract __scores__ (typically numeric). The most natural example is sport results, however not the only one. For example, product rating can be considered as a competition between products as "players". Here a "game" is a customer that reviews a set of products by rating them with numerical "score" (stars, points, etc.).

We will need the following packages:

```{r library, warning = FALSE}
library(comperes)
library(tibble)
```

## Storage

### Long format

Results in long format are stored in object of class `longcr`. It is considered to be a `tibble` with one row per game-player pair. It should have at least columns with names "game", "player" and "score". For example:

```{r cr_long_raw}
cr_long_raw <- tibble(
  game   = c(1,  1,  1, 2, 2, 3, 3, 4),
  player = c(1, NA, NA, 1, 2, 2, 1, 2),
  score  = 1:8
)
```

To convert `cr_long_raw` into `longcr` object use `as_longcr()`:

```{r cr_long}
cr_long <- as_longcr(cr_long_raw)
cr_long
```

By default, `as_longcr()` repairs its input by applying set of heuristics to extract relevant data:

```{r as_longcr-repair}
tibble(
  PlayerRS = "a",
  gameSS = "b",
  extra = -1,
  score_game = 10,
  player = 1
) %>%
  as_longcr()
```

### Wide format

Results in wide format are stored in object of class `widecr`. It is considered to be a `tibble` with one row per game with fixed amount of players. Data should be organized in pairs of columns "player"-"score". Identifier of a pair should go after respective keyword and consist only from digits. For example: player1, score1, player2, score2. Order doesn't matter.

Extra columns are allowed. Column game for game identifier is optional.

Example of correct wide format:

```{r cr_wide_raw}
cr_wide_raw <- tibble(
  player1 = c(1, 1, 2),
  score1  = -(1:3),
  player2 = c(2, 3, 3),
  score2  = -(4:6)
)
```

To convert `cr_wide_raw` into `widecr` object use `as_widecr()`:

```{r cr_wide}
cr_wide <- cr_wide_raw %>% as_widecr()
cr_wide
```

By default, `as_widecr()` also does repairing of its input:

```{r as_widecr-repair}
tibble(
  score = 2,
  PlayerRS = "a",
  scoreRS = 1,
  player = "b",
  player1 = "c",
  extra = -1,
  game = "game"
) %>%
  as_widecr()
```

## Conversion

`as_longcr()` and `as_widecr()` do actual conversion applied to `widecr` and `longcr` objects respectively:

```{r conversion}
as_longcr(cr_wide)

# Determines number of players in game as
# actual maximum number of players in games
as_widecr(cr_long)
```

## Notes

- Functions in `comperes` expect data that can be a proper input to `as_longcr()`, i.e. `longcr` object, `widecr` object, or raw data aligned with long format.
- The preferred way to do data analysis with `comperes` is to have three data frames:
    - One with description of __games__ (with column `game` for game identifiers).
    - One with description of __players__ (with column `player` for player identifiers).
    - One with __competition results__ in long format.
    
    This way one can operate with games between variable number of players with minimum storage overhead.