| Type: | Package |
| Title: | Acute Chronic Workload Ratio Calculation |
| Version: | 0.2.0 |
| Maintainer: | Jorge R Fernandez-Santos <jorgedelrosario.fernandez@uca.es> |
| Description: | Functions for calculating the acute chronic workload ratio using three different methods: exponentially weighted moving average (EWMA), rolling average coupled (RAC) and rolling averaged uncoupled (RAU). Sessions are placed on a daily calendar, so rest days count as zero load. Examples of these methods can be found in Williams et al. (2017) <doi:10.1136/bjsports-2016-096589> for EWMA and Windt & Gabbett (2019) <doi:10.1136/bjsports-2017-098925> for RAC and RAU. |
| License: | MIT + file LICENSE |
| Encoding: | UTF-8 |
| LazyData: | true |
| Imports: | r2d3 |
| Depends: | R (≥ 2.10) |
| RoxygenNote: | 7.3.3 |
| URL: | https://jorgedelro.github.io/ACWR/, https://github.com/JorgeDelro/ACWR |
| BugReports: | https://github.com/JorgeDelro/ACWR/issues |
| Suggests: | knitr, rmarkdown, testthat (≥ 3.0.0) |
| Config/testthat/edition: | 3 |
| VignetteBuilder: | knitr |
| NeedsCompilation: | no |
| Packaged: | 2026-11-17 08:38:35 UTC; jorge |
| Author: | Jorge R Fernandez-Santos
|
| Repository: | CRAN |
| Date/Publication: | 2026-11-17 10:30:18 UTC |
Acute Chronic Workload Ratio
Description
Calculates the acute chronic workload ratio for one or several athletes
with one or more methods (see EWMA(), RAC() and RAU()).
Usage
ACWR(
db,
ID,
TL,
weeks = NULL,
days = NULL,
training_dates,
ACWR_method = c("EWMA", "RAC", "RAU"),
partial = FALSE
)
Arguments
db |
a data frame with one row per session. |
ID |
name of the column with the ID of the subjects. |
TL |
name of the column with the training load (numeric, >= 0). |
weeks |
deprecated and ignored in the calculation. If given, the column is kept in the output. |
days |
deprecated and ignored in the calculation. If given, the
column is kept in the output (e.g. for |
training_dates |
name of the column with the training dates (Date or POSIXct). Rows do not need to be sorted. |
ACWR_method |
method(s) to calculate ACWR: any of "EWMA", "RAC" and "RAU". |
partial |
passed to |
Details
Each athlete's sessions are turned into a daily series from their first to their last date: sessions on the same day are summed and days without a session count as a load of 0.
A missing load (NA) makes every window that contains its day NA, and a single warning is given for the whole data frame.
Value
a data frame with the ID, TL, weeks, days and training dates
columns, followed by the acute & chronic training load and ACWR of
the selected method/s. Rows are in the same order as in db.
Package decisions
The following choices are made by this package; they are not taken from the cited sources:
RAC and RAU windows move day by day. The sources define acute and chronic loads by weeks (the most recent week, and the average of the last 4 weeks or of the 3 preceding weeks). This package interprets them as rolling windows of 7, 28 and 21 calendar days that end on each day, so every day has its own value.
With
partial = TRUE, a window that starts before the first day uses only the days available and is rescaled to 7 days (sum x 7 / number of days available). The sources do not cover incomplete windows.For all methods, the ratio is NA when the chronic load is 0 (e.g. no training in the chronic window): a ratio to zero load is undefined, and returning NA avoids infinite values in summaries and plots.
EWMA is updated every calendar day, as defined by Williams et al. (2017); this is not a package interpretation.
Examples
data("training_load", package = "ACWR")
# Calculate ACWR
result_ACWR <- ACWR(db = training_load,
ID = "ID",
TL = "TL",
training_dates = "Training_Date",
ACWR_method = c("EWMA", "RAC", "RAU"))
# Athletes with rest days
data("training_load_rest", package = "ACWR")
result_rest <- ACWR(db = training_load_rest,
ID = "ID",
TL = "TL",
training_dates = "Training_Date",
ACWR_method = c("EWMA", "RAC", "RAU"))
Exponentially Weighted Moving Average
Description
Calculates the acute chronic workload ratio with exponentially weighted moving averages (Williams et al., 2017).
Usage
EWMA(TL, training_dates = NULL)
Arguments
TL |
training load of each session (numeric, >= 0). |
training_dates |
date of each session (Date or POSIXct). Sessions do
not need to be sorted. If |
Details
Sessions are first turned into a daily series from the first to the last
date: sessions on the same day are summed and days without a session count
as a load of 0. Then, for every calendar day t with daily load
w_t:
E_t = w_t \lambda + (1 - \lambda) E_{t-1}, \quad \lambda = 2 / (N + 1)
with N = 7 (acute) and N = 28 (chronic), and E_1 = w_1.
Acute and chronic loads are on the scale of a daily load. Each session gets
the value of its day. Values are not rounded.
EWMA is recursive: a missing load (NA) makes that day and every later day NA. A warning is given; impute missing loads before calling the function.
Value
A list with one value per session, in the input order:
EWMA_acute: EWMA - acute training load.
EWMA_chronic: EWMA - chronic training load.
EWMA_ACWR: EWMA - Acute-Chronic Workload Ratio (NA when the chronic load is 0).
Package decisions
The following choices are made by this package; they are not taken from the cited sources:
The ratio is NA when the chronic load is 0 (e.g. no training in the chronic window): a ratio to zero load is undefined, and returning NA avoids infinite values in summaries and plots.
Updating the EWMA every calendar day, including rest days, follows the definition of Williams et al. (2017) and is not a package interpretation.
References
Williams S, West S, Cross MJ, Stokes KA (2017). Better way to determine the acute:chronic workload ratio? British Journal of Sports Medicine, 51(3), 209-210. doi:10.1136/bjsports-2016-096589
Examples
data("training_load", package = "ACWR")
# Select the first subject
training_load_1 <- training_load[training_load[["ID"]] == 1, ]
# Calculate ACWR
result_EWMA <- EWMA(TL = training_load_1$TL,
training_dates = training_load_1$Training_Date)
Rolling Average Coupled
Description
Calculates the acute chronic workload ratio with rolling averages, where the acute week is included in the chronic period (coupled ratio).
Usage
RAC(TL, weeks = NULL, training_dates, partial = FALSE)
Arguments
TL |
training load of each session (numeric, >= 0). |
weeks |
deprecated and ignored. Kept for backwards compatibility;
the calculation now uses |
training_dates |
date of each session (Date or POSIXct). Sessions do not need to be sorted. |
partial |
if |
Details
Sessions are first turned into a daily series from the first to the last
date: sessions on the same day are summed and days without a session count
as a load of 0. Then, for each day t:
Acute load: sum of the daily load over days
t-6tot.Chronic load: sum of the daily load over days
t-27tot, divided by 4 (weekly-equivalent).ACWR: acute / chronic.
Each session gets the value of its day. Values are not rounded.
A missing load (NA) makes every window that contains its day NA, and a warning is given.
Value
A list with one value per session, in the input order:
RAC_acute: RAC - acute training load.
RAC_chronic: RAC - chronic training load.
RAC_ACWR: RAC - Acute-Chronic Workload Ratio (NA when the chronic load is 0).
Package decisions
The following choices are made by this package; they are not taken from the cited sources:
Windows move day by day. The sources define acute and chronic loads by weeks (the most recent week, and the average of the last 4 weeks or of the 3 preceding weeks). This package interprets them as rolling windows of 7, 28 and 21 calendar days that end on each day, so every day has its own value.
With
partial = TRUE, a window that starts before the first day uses only the days available and is rescaled to 7 days (sum x 7 / number of days available). The sources do not cover incomplete windows.The ratio is NA when the chronic load is 0 (e.g. no training in the chronic window): a ratio to zero load is undefined, and returning NA avoids infinite values in summaries and plots.
References
Windt J, Gabbett TJ (2019). Is it all for naught? What does mathematical coupling mean for acute:chronic workload ratios? British Journal of Sports Medicine, 53(16), 988-990. doi:10.1136/bjsports-2017-098925
Examples
data("training_load", package = "ACWR")
# Select the first subject
training_load_1 <- training_load[training_load[["ID"]] == 1, ]
# Calculate ACWR
result_RAC <- RAC(TL = training_load_1$TL,
training_dates = training_load_1$Training_Date)
Rolling Average Uncoupled
Description
Calculates the acute chronic workload ratio with rolling averages, where the acute week is excluded from the chronic period (uncoupled ratio).
Usage
RAU(TL, weeks = NULL, training_dates, partial = FALSE)
Arguments
TL |
training load of each session (numeric, >= 0). |
weeks |
deprecated and ignored. Kept for backwards compatibility;
the calculation now uses |
training_dates |
date of each session (Date or POSIXct). Sessions do not need to be sorted. |
partial |
if |
Details
Sessions are first turned into a daily series from the first to the last
date: sessions on the same day are summed and days without a session count
as a load of 0. Then, for each day t:
Acute load: sum of the daily load over days
t-6tot.Chronic load: sum of the daily load over days
t-27tot-7, divided by 3 (weekly-equivalent).ACWR: acute / chronic.
Each session gets the value of its day. Values are not rounded.
A missing load (NA) makes every window that contains its day NA, and a warning is given.
Value
A list with one value per session, in the input order:
RAU_acute: RAU - acute training load.
RAU_chronic: RAU - chronic training load.
RAU_ACWR: RAU - Acute-Chronic Workload Ratio (NA when the chronic load is 0).
Package decisions
The following choices are made by this package; they are not taken from the cited sources:
Windows move day by day. The sources define acute and chronic loads by weeks (the most recent week, and the average of the last 4 weeks or of the 3 preceding weeks). This package interprets them as rolling windows of 7, 28 and 21 calendar days that end on each day, so every day has its own value.
With
partial = TRUE, a window that starts before the first day uses only the days available and is rescaled to 7 days (sum x 7 / number of days available). The sources do not cover incomplete windows.The ratio is NA when the chronic load is 0 (e.g. no training in the chronic window): a ratio to zero load is undefined, and returning NA avoids infinite values in summaries and plots.
References
Windt J, Gabbett TJ (2019). Is it all for naught? What does mathematical coupling mean for acute:chronic workload ratios? British Journal of Sports Medicine, 53(16), 988-990. doi:10.1136/bjsports-2017-098925
Examples
data("training_load", package = "ACWR")
# Select the first subject
training_load_1 <- training_load[training_load[["ID"]] == 1, ]
# Calculate ACWR
result_RAU <- RAU(TL = training_load_1$TL,
training_dates = training_load_1$Training_Date)
ACWR plots using d3.js
Description
ACWR plots using d3.js
Usage
plot_ACWR(
db,
TL,
ACWR,
day,
ID = NULL,
colour = NULL,
xLabel = NULL,
y0Label = NULL,
y1Label = NULL,
plotTitle = NULL
)
Arguments
db |
a data frame |
TL |
training load |
ACWR |
Acute Chronic Workload Ratio |
day |
training days: day numbers (whole numbers) or dates (Date or POSIXct). The x-axis is a continuous calendar from the first to the last day. |
ID |
ID of the subjects |
colour |
colour of the bars. By default "#87CEEB" (skyblue) |
xLabel |
x-axis label. By default "Days" |
y0Label |
left y-axis label. By default |
y1Label |
right y-axis label. By default "Acute:chronic workload ratio" |
plotTitle |
Title of the plot. By default "ACWR" |
Value
This function returns a d3.js object for a single subject. For several subjects it returns a list of d3.js objects. The x-axis has one slot per calendar day between the first and the last day, labelled every 7 days; days without a session have a bar of height 0. Sessions on the same day are shown as one bar (summed load). The ACWR line joins the sessions that have a value, across rest days, and is not drawn where the ACWR is NA.
Examples
data("training_load", package = "ACWR")
# Calculate ACWR
result_ACWR <- ACWR(db = training_load,
ID = "ID",
TL = "TL",
days = "Day",
training_dates = "Training_Date",
ACWR_method = c("EWMA", "RAC", "RAU"))
# Plot for 1 subject
# Select the first subject
result_ACWR_1 <- result_ACWR[result_ACWR[["ID"]] == 1, ]
# plot ACWR (e.g. EWMA)
ACWR_plot_1 <- plot_ACWR(db = result_ACWR_1,
TL = "TL",
ACWR = "EWMA_ACWR",
day = "Day")
# Plot for several subjects
# plot ACWR (e.g. RAC)
ACWR_plot <- plot_ACWR(db = result_ACWR,
TL = "TL",
ACWR = "RAC_ACWR",
day = "Day",
ID = "ID")
Training load dataframe
Description
A dataframe with the training load of 3 subjects.
Usage
data("training_load", package = "ACWR")
Format
An object of class tbl_df (inherits from tbl, data.frame) with 84 rows and 5 columns.
Variables
- ID
ID of the subjects
- Week
training weeks
- Day
training days
- TL
training load (arbitrary units)
- Training_Date
training dates
Training load dataframe with rest days
Description
A dataframe with the training load of 3 subjects over 8 weeks,
with rest days. Subject 1 trains 4 days per week; subject 2 trains 5 days
per week with a double session on Wednesdays and does not train in week 5;
subject 3 trains 3 days per week and 6 days in the last week. Created with
simulated data (see data-raw/training_load_rest.R in the source code).
Usage
data("training_load_rest", package = "ACWR")
Format
A data frame with 101 rows (one per session) and 5 variables:
- ID
ID of the subjects
- Week
training week
- Day
calendar day since the start (1 = Monday 3 January 2022)
- TL
training load (arbitrary units)
- Training_Date
training date (Date)