As said in my title, I’ve got a database of fighters’ statistics where the names alternate. My database illustrates what I mean :

Database <- c("Fighter A", "Fighter B", "Fighter A", "Fighter  B",
                   "Fighter A", "Fighter B", "Fighter A", "Fighter B",
                   "Fighter  C", "Fighter  D", "Fighter  C", "Fighter  D", 
                   "Fighter  A", "Fighter C")

I want to add a column based on the observations in the fighters’ names column. In my example above I would like a round column built as follows:

Database <- mutate(round = c(0, 0, 1, 1, 2, 2, 3, 3, 0, 0, 1, 1, 0, 0)

If the pair (fighter A and fighter B for example) is repeated, the round will receive 0, 0, 1, 1 as many times as the pattern is repeated. If the pattern is broken (fighter C and fighter D) then the round restarts at 0. It’s a way to count the number of rounds where round = 0 is the average of the following rounds. In all cases, the round will always have 0, 0, 1, 1 values.

Here’s what I tried without success:

fighters_name <- c("Combattant A", "Combattant B", "Combattant A", "Combattant B",
                   "Combattant A", "Combattant B", "Combattant A", "Combattant B",
                   "Combattant C", "Combattant D", "Combattant C", "Combattant D", 
                   "Combattant A", "Combattant C")


df <- tibble(fighters_name)

df <- df %>%
  mutate(pair_id = cumsum(fighters_name != lag(fighters_name, default = fighters_name[1]))) %>%
  group_by(pair_id) %>%
  mutate(round = (row_number() - 1) %% 2) %>%
  ungroup() %>%
  select(-pair_id)

If someone has an idea to build the round column I would be grateful. Thank you.