Compressing/combining rows in r df based on multiple conditions

I have a df containing the outputs of a simple game. Every 30 rows, the df switches to a new set of players. Here is a sample with just the first 2 sets of players:

> dput(stackdf)
structure(list(id = structure(list(`dataset_single$id` = c(20230420, 
20230420, 20230420, 20230420, 20230420, 20230420, 20230420, 20230420, 
20230420, 20230420, 20230420, 20230420, 20230420, 20230420, 20230420, 
20230420, 20230420, 20230420, 20230420, 20230420, 20230420, 20230420, 
20230420, 20230420, 20230420, 20230420, 20230420, 20230420, 20230420, 
20230420, 20230424, 20230424, 20230424, 20230424, 20230424, 20230424, 
20230424, 20230424, 20230424, 20230424, 20230424, 20230424, 20230424, 
20230424, 20230424, 20230424, 20230424, 20230424, 20230424, 20230424, 
20230424, 20230424, 20230424, 20230424, 20230424, 20230424, 20230424, 
20230424, 20230424, 20230424)), row.names = c(NA, -60L), class = "data.frame"), 
round = c("1", "2", "3", "4", "5", "6", "7", "8", "9", "10", 
"11", "12", "13", "14", "15", "16", "17", "18", "19", "20", 
"21", "22", "23", "24", "25", "26", "27", "28", "29", "30", 
"1", "2", "3", "4", "5", "6", "7", "8", "9", "10", "11", 
"12", "13", "14", "15", "16", "17", "18", "19", "20", "21", 
"22", "23", "24", "25", "26", "27", "28", "29", "30"), win = c(1, 
0, 1, 1, 1, 1, 0, 0, 1, 0, 0, 1, 1, 1, 1, 0, 1, 0, 1, 0, 
1, 0, 1, 1, 1, 1, 0, 0, 0, 0, 0, 1, 1, 0, 1, 0, 1, 0, 1, 
1, 0, 0, 1, 1, 1, 0, 1, 1, 1, 1, 1, 0, 1, 0, 0, 1, 1, 0, 
0, 0), sound = c(1, 1, 2, 1, 2, 1, 1, 2, 2, 2, 1, 1, 2, 3, 
5, 6, 7, 8, 8, 1, 4, 8, 6, 7, 8, 5, 4, 8, 8, 8, 3, 4, 5, 
5, 6, 5, 2, 5, 7, 4, 7, 6, 8, 5, 1, 8, 5, 3, 7, 6, 3, 6, 
6, 5, 5, 8, 1, 6, 6, 5)), row.names = c(NA, -60L), class = c("tbl_df", 
"tbl", "data.frame"))

The "win" column indicates the winner (p1=1, p2=0), and the "sound" column indicates the score of the winner for that round. As you can see, there are streaks where p1 might win a few rounds in a row, and then p2 wins several, etc. When there is a streak, I want to compress the streak so that it becomes one row, where the "sound" column takes a mean of the sound scores in that streak. For example, this:

would become:

There are 2 things making this tricky: 1.) the full df is huge, so I am seeking an automated solution (probably a for loop?); and 2.) I need to confine the program to one set of players at a time so that it does not think that rows 27:31 (for example) is one big streak (as row 31 is actually a new set of players).

Any help would be great

It seems the first id column is a nested dataframe inside a column. I have corrected it to a regular column.

stackdf$id <- stackdf$id$`dataset_single$id`

Here are two option using dplyr and data.table which follows the same logic. Use data.table option if you need more speed.

dplyr -

library(dplyr)
stackdf %>%
group_by(id, cons_id = consecutive_id(win)) %>% 
summarise(win = first(win), 
sound = mean(sound),.groups = "drop")

data.table -

library(data.table)  
setDT(stackdf)
stackdf[, .(win = first(win), sound = mean(sound)), .(id, cons_id = rleid(win))]

consecutive_id in dplyr and rleid in data.table creates a unique identifier which increments every time the value changes.

Finally it looks as a group and summarise task:
Note the first column looks like a nested list:

library(tidyr)
library(dplyr) #> 1.1.0
stackdf %>%
unnest(id) %>% 
rename(id = `dataset_single$id`) %>% 
mutate(x = consecutive_id(win), .by=id) %>% 
summarise(mean_sound = mean(sound), .by = c(id, x, win)) %>% 
select(-x)

  id   win mean_sound
<dbl> <dbl>      <dbl>
1 20230420     1       1   
2 20230420     0       1   
3 20230420     1       1.5 
4 20230420     0       1.5 
5 20230420     1       2   
6 20230420     0       1.5 
7 20230420     1       2.75
8 20230420     0       6   
9 20230420     1       7   
10 20230420     0       8   
# … with 23 more rows
# ℹ Use `print(n = ...)` to see more rows