Notes: - Models are ONLY trained on Jokic career data from the present date to the 2019 season, model will ideally be trained on all players that are currently active. - H and F are currently manually selected.

Packages

#Packages
library(tidyverse)
library(jsonlite)
library(janitor)
library(dplyr)
library(DEoptim)
library(Metrics)
library(tune)
library(dlm)
library(xgboost)

Data pull and manipulation

pbp_jok_url <- "https://api.pbpstats.com/get-game-logs/nba?Season=2023-24,2022-23,2021-22,2020-21,2019-20,2018-19&SeasonType=Regular%2BSeason&EntityId=203999&EntityType=Player"
pbp_jok_json <- read_json(pbp_jok_url)
pbp_jok <- pbp_jok_json[["multi_row_table_data"]] %>%
  bind_rows() %>%
  clean_names()

pbp_jok_play_url <- "https://api.pbpstats.com/get-game-logs/nba?Season=2023-24,2022-23,2021-22,2020-21,2019-20,2018-19&SeasonType=Playoffs&EntityId=203999&EntityType=Player"
pbp_jok_play_json <- read_json(pbp_jok_play_url)
pbp_jok_play <- pbp_jok_play_json[["multi_row_table_data"]] %>%
  bind_rows() %>%
  clean_names()

# fix weights so that when each observation is calculated its adjusted

today <- Sys.Date()

days <- difftime(today, pbp_jok$date, unit = "days") %>%
  as.data.frame() %>%
  mutate_at(vars(.), funs(round(., 0)))
days$. <- as.numeric(days$.)
days <- days %>%
  rename(days = '.')

jok_w_days <- cbind(pbp_jok, days) %>%
  replace(is.na(.), 0)

Exponential Weighting

jok_3pt <- jok_w_days %>%
  filter(fg3a > 0) %>%
  mutate(avg_3pt_pct = mean(fg3pct))

beta_3pt <- function(x){
  rmse(jok_3pt$avg_3pt_pct, (x^(exp(jok_3pt$days))*jok_3pt$fg3pct))
}

lower <- c(0)
upper <- c(.99)

set.seed(188)
B_3pt <- DEoptim(beta_3pt , lower > 0, upper < 1, DEoptim.control(itermax = 5))

## Iteration: 1 bestvalit: 0.338873 bestmemit:    0.914971
## Iteration: 2 bestvalit: 0.338691 bestmemit:    0.966314
## Iteration: 3 bestvalit: 0.338691 bestmemit:    0.966314
## Iteration: 4 bestvalit: 0.338409 bestmemit:    0.991148
## Iteration: 5 bestvalit: 0.338409 bestmemit:    0.991148

jok_3pt <- jok_3pt %>%
  mutate( fg3pct_game_estimate = jok_3pt$fg3pct * (B_3pt$optim$bestmem^(expo_decay(jok_3pt$days, start_val = 0, limit_val = 5)))) %>%
  mutate(fg3pct_running_avg = cumsum(fg3pct_game_estimate) / (1:n()))

jok_2pt <- jok_w_days %>%
  filter(fg2a > 0) %>%
   mutate(avg_2pt_pct = mean(fg2pct))

beta_2pt <- function(x){
  rmse(jok_2pt$avg_2pt_pct, (x^(exp(jok_2pt$days))*jok_2pt$fg2pct))
}

lower <- c(0)
upper <- c(.99)

set.seed(1986)
B_2pt <- DEoptim(beta_2pt , lower > 0, upper < 1, DEoptim.control(itermax = 5))

## Iteration: 1 bestvalit: 0.625306 bestmemit:    0.934639
## Iteration: 2 bestvalit: 0.625306 bestmemit:    0.934639
## Iteration: 3 bestvalit: 0.624756 bestmemit:    0.982517
## Iteration: 4 bestvalit: 0.624287 bestmemit:    0.994284
## Iteration: 5 bestvalit: 0.624287 bestmemit:    0.994284

jok_2pt <- jok_2pt %>%
  mutate( fg2pct_game_estimate = jok_2pt$fg2pct * (B_2pt$optim$bestmem^(expo_decay(jok_2pt$days, start_val = 0, limit_val = 5)))) %>%
  mutate(fg2pct_running_avg = cumsum(fg2pct_game_estimate) / (1:n()))

jok_assist <- jok_w_days %>%
  filter(assists > 0) %>%
   mutate(avg_assists = mean(assists))

beta_assist <- function(x){
  rmse(jok_assist$avg_assists, (x^(exp(jok_assist$days))*jok_assist$assists))
}

lower <- c(0)
upper <- c(.99)

set.seed(194)
B_assist <- DEoptim(beta_assist , lower > 0, upper < 1, DEoptim.control(itermax = 5))

## Iteration: 1 bestvalit: 8.128287 bestmemit:    0.968275
## Iteration: 2 bestvalit: 8.128287 bestmemit:    0.968275
## Iteration: 3 bestvalit: 8.122512 bestmemit:    0.987329
## Iteration: 4 bestvalit: 8.122512 bestmemit:    0.987329
## Iteration: 5 bestvalit: 8.115084 bestmemit:    0.997644

jok_assist <- jok_assist %>%
  mutate( assist_game_estimate = jok_assist$assists * (B_assist$optim$bestmem^(expo_decay(jok_assist$days, start_val = 0, limit_val = 5)))) %>%
  mutate(assists_running_avg = cumsum(assist_game_estimate) / (1:n()))

Kalman Filter

#function for initialization
initializeKalmanFilter <- function(initial_state, initial_covariance, process_noise, measurement_noise){
  kf <- list()
  kf$state <- initial_state
  kf$covariance <- initial_covariance
  kf$process_noise <- process_noise
  kf$measurement_noise <- measurement_noise
  
  return(kf)
}

#function for prediction
predictKalmanFilter <- function(kf,F){
  predicted_state <- F * kf$state
  predicted_covariance <- F * kf$covariance * F + kf$process_noise
  
  return(list(predicted_state = predicted_state, predicted_covariance = predicted_covariance))
}

#function for updating
updateKalmanFilter <- function(kf, predicted_state, predicted_covariance, measurement, H){
  K <- predicted_covariance * H / (H * predicted_covariance * H + kf$measurement_noise)
  updated_state <- predicted_state + K * (measurement - H * predicted_state)
  updated_covariance <- (1 - K * H) * predicted_covariance
  
  return(list(updated_state = updated_state , updated_covariance = updated_covariance))
}

#function to perform Kalman filtering
kalmanFilter <- function( initial_state, initial_covariance, process_noise, measurement_noise, measurements, F, H){
  kf <- initializeKalmanFilter(initial_state, initial_covariance, process_noise, measurement_noise)
  filtered_states <- vector(mode = "numeric", length = length(measurements))
  
  for (i in 1:length(measurements)){
    prediction <- predictKalmanFilter(kf, F)
    update <- updateKalmanFilter(kf, prediction$predicted_state, prediction$predicted_covariance, measurements[i], H)
    kf$state <- update$updated_state
    kf$covariance <- update$updated_covariance
    filtered_states[i] <- kf$state
  }
  
  return(filtered_states)
}

# 3pt% run thru Kalman filter
initial_state <- .326
initial_covariance <- 1
measurement_noise <- jok_3pt$fg3a/244
process_noise <- var(jok_3pt$fg3pct_game_estimate)
measurements <- jok_3pt$fg3pct
F <- .5
H <- .3

kalman_3pt <- kalmanFilter(initial_state, initial_covariance, process_noise, measurement_noise, measurements, F, H)

kalman_3pt <- kalman_3pt %>%
  as.data.frame() %>%
  rename(kalman_3pt_pct = '.')

# 2pt% run thru Kalman filter
initial_state <- .600
initial_covariance <- 1
measurement_noise <- jok_2pt$fg2a/127
process_noise <- var(jok_2pt$fg2pct_game_estimate)
measurements <- jok_2pt$fg2pct
F <- .9
H <- .75

kalman_2pt <- kalmanFilter(initial_state, initial_covariance, process_noise, measurement_noise, measurements, F, H)

kalman_2pt <- kalman_2pt %>%
  as.data.frame() %>%
  rename(kalman_2pt_pct = '.')

Gradient Boosted Decision Tree

#3pt_pct combined with gradient boosted decision tree

fg3pt_pct_pred <- cbind(jok_3pt, kalman_3pt) %>%
  select(kalman_3pt_pct,
         fg3pct_running_avg,
         avg_3pt_pct)
xg_train <- xgb.DMatrix(data = as.matrix(fg3pt_pct_pred), label = fg3pt_pct_pred$avg_3pt_pct)
x_model <- xgboost( data = xg_train, max.depth = 3 , eta = .2 , nrounds = 50, objective = "binary:logistic")

## [1]  train-logloss:0.674642 
## [2]  train-logloss:0.662673 
## [3]  train-logloss:0.654884 
## [4]  train-logloss:0.649801 
## [5]  train-logloss:0.646482 
## [6]  train-logloss:0.644317 
## [7]  train-logloss:0.642906 
## [8]  train-logloss:0.641987 
## [9]  train-logloss:0.641390 
## [10] train-logloss:0.641003 
## [11] train-logloss:0.640752 
## [12] train-logloss:0.640589 
## [13] train-logloss:0.640484 
## [14] train-logloss:0.640416 
## [15] train-logloss:0.640372 
## [16] train-logloss:0.640344 
## [17] train-logloss:0.640325 
## [18] train-logloss:0.640314 
## [19] train-logloss:0.640306 
## [20] train-logloss:0.640301 
## [21] train-logloss:0.640298 
## [22] train-logloss:0.640296 
## [23] train-logloss:0.640295 
## [24] train-logloss:0.640294 
## [25] train-logloss:0.640293 
## [26] train-logloss:0.640293 
## [27] train-logloss:0.640293 
## [28] train-logloss:0.640293 
## [29] train-logloss:0.640292 
## [30] train-logloss:0.640292 
## [31] train-logloss:0.640292 
## [32] train-logloss:0.640292 
## [33] train-logloss:0.640292 
## [34] train-logloss:0.640292 
## [35] train-logloss:0.640292 
## [36] train-logloss:0.640292 
## [37] train-logloss:0.640292 
## [38] train-logloss:0.640292 
## [39] train-logloss:0.640292 
## [40] train-logloss:0.640292 
## [41] train-logloss:0.640292 
## [42] train-logloss:0.640292 
## [43] train-logloss:0.640292 
## [44] train-logloss:0.640292 
## [45] train-logloss:0.640292 
## [46] train-logloss:0.640292 
## [47] train-logloss:0.640292 
## [48] train-logloss:0.640292 
## [49] train-logloss:0.640292 
## [50] train-logloss:0.640292

predict(x_model, as.matrix(fg3pt_pct_pred))

##   [1] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
##   [8] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
##  [15] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
##  [22] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
##  [29] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
##  [36] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
##  [43] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
##  [50] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
##  [57] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
##  [64] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
##  [71] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
##  [78] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
##  [85] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
##  [92] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
##  [99] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
## [106] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
## [113] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
## [120] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
## [127] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
## [134] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
## [141] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
## [148] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
## [155] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
## [162] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
## [169] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
## [176] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
## [183] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
## [190] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
## [197] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
## [204] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
## [211] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
## [218] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
## [225] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
## [232] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
## [239] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
## [246] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
## [253] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
## [260] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
## [267] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
## [274] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
## [281] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
## [288] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
## [295] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
## [302] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
## [309] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
## [316] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
## [323] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
## [330] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
## [337] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
## [344] 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865 0.3388865
## [351] 0.3388865 0.3388865 0.3388865 0.3388865

#2pt_pct combined with gradient boosted decision tree

fg2pt_pct_pred <- cbind(jok_2pt, kalman_2pt) %>%
  select(kalman_2pt_pct,
         fg2pct_running_avg,
         avg_2pt_pct)
xg_train_2pt <- xgb.DMatrix(data = as.matrix(fg2pt_pct_pred), label = fg2pt_pct_pred$avg_2pt_pct)
x_model_2pt <- xgboost( data = xg_train_2pt, max.depth = 3 , eta = .2 , nrounds = 50, objective = "binary:logistic")

## [1]  train-logloss:0.681843 
## [2]  train-logloss:0.674550 
## [3]  train-logloss:0.669827 
## [4]  train-logloss:0.666763 
## [5]  train-logloss:0.664775 
## [6]  train-logloss:0.663486 
## [7]  train-logloss:0.662650 
## [8]  train-logloss:0.662109 
## [9]  train-logloss:0.661759 
## [10] train-logloss:0.661533 
## [11] train-logloss:0.661386 
## [12] train-logloss:0.661292 
## [13] train-logloss:0.661231 
## [14] train-logloss:0.661192 
## [15] train-logloss:0.661167 
## [16] train-logloss:0.661150 
## [17] train-logloss:0.661140 
## [18] train-logloss:0.661133 
## [19] train-logloss:0.661129 
## [20] train-logloss:0.661126 
## [21] train-logloss:0.661124 
## [22] train-logloss:0.661123 
## [23] train-logloss:0.661122 
## [24] train-logloss:0.661122 
## [25] train-logloss:0.661121 
## [26] train-logloss:0.661121 
## [27] train-logloss:0.661121 
## [28] train-logloss:0.661121 
## [29] train-logloss:0.661121 
## [30] train-logloss:0.661121 
## [31] train-logloss:0.661121 
## [32] train-logloss:0.661121 
## [33] train-logloss:0.661121 
## [34] train-logloss:0.661121 
## [35] train-logloss:0.661121 
## [36] train-logloss:0.661121 
## [37] train-logloss:0.661121 
## [38] train-logloss:0.661121 
## [39] train-logloss:0.661121 
## [40] train-logloss:0.661121 
## [41] train-logloss:0.661121 
## [42] train-logloss:0.661121 
## [43] train-logloss:0.661121 
## [44] train-logloss:0.661121 
## [45] train-logloss:0.661121 
## [46] train-logloss:0.661121 
## [47] train-logloss:0.661121 
## [48] train-logloss:0.661121 
## [49] train-logloss:0.661121 
## [50] train-logloss:0.661121

predict(x_model_2pt, as.matrix(fg2pt_pct_pred))

##   [1] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
##   [8] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
##  [15] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
##  [22] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
##  [29] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
##  [36] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
##  [43] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
##  [50] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
##  [57] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
##  [64] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
##  [71] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
##  [78] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
##  [85] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
##  [92] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
##  [99] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
## [106] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
## [113] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
## [120] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
## [127] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
## [134] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
## [141] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
## [148] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
## [155] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
## [162] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
## [169] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
## [176] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
## [183] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
## [190] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
## [197] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
## [204] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
## [211] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
## [218] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
## [225] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
## [232] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
## [239] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
## [246] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
## [253] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
## [260] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
## [267] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
## [274] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
## [281] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
## [288] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
## [295] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
## [302] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
## [309] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
## [316] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
## [323] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
## [330] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
## [337] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
## [344] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
## [351] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
## [358] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
## [365] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
## [372] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
## [379] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
## [386] 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606 0.6258606
## [393] 0.6258606 0.6258606

```

3pt Shooting projection

Dylan Kinion

202311-26

Packages

Data pull and manipulation

Exponential Weighting

Kalman Filter

Gradient Boosted Decision Tree