gistfile1.txt

``` r
library(slider)
library(tidyverse)

set.seed(20210715)

sim_data <- crossing(
  player = letters[1:7],
  year = 2019:2020,
  week = 1:17
) %>% 
  mutate(
    row_n = row_number(),
    qbr = rnorm(n = max(row_n), mean = 70, sd = 8)
  )

my_roll_mean <- function(x){
  weighted.mean(x, w = c(.1, 0.25, 0.5, 0.75, 1))
}

plot_data <- sim_data %>% 
  group_by(player) %>% 
  mutate(
    pure_roll_avg = slide_dbl(qbr, mean, .before = 4L, .complete = TRUE),
    weight_roll_avg = slide_dbl(qbr, my_roll_mean, .before = 4L, .complete = TRUE)
  ) %>% 
  mutate(grp_row = row_number()) %>% 
  ungroup() %>% 
  nest_by(player, .keep = TRUE) %>% 
  summarize(data = bind_cols(
    data, 
    loess = loess(qbr ~ grp_row, degree = 1, span = 0.2, data = data)$fitted),
    .groups = "drop"
  ) %>% 
  pluck("data")

plot_data %>% 
  mutate(loess = if_else(is.na(weight_roll_avg), NA_real_, loess)) %>% 
  pivot_longer(cols=c(pure_roll_avg, weight_roll_avg, loess), names_to = "type", values_to = "avg") %>% 
  ggplot(aes(x = grp_row, y = avg, color = type)) +
  geom_line() +
  facet_wrap(~player)
#> Warning: Removed 12 row(s) containing missing values (geom_path).
```

![](https://i.imgur.com/I8TJzVP.png)

<sup>Created on 2021-10-25 by the [reprex package](https://reprex.tidyverse.org) (v2.0.1)</sup>