Hi, first thanks for this great package, it is very interesting and nicely designed!
I ran into the issue with elo.run.multiteam where it ignores intial ratings so for now I rewrote the function to pass it to elo.run (not sure it s the correct way but looks ok)
elo.run.multiteam2 <- function (formula, data, na.action, subset, k = NULL, initial.elos = NULL,
...)
{
Call <- match.call()
Call[[1L]] <- quote(elo::elo.model.frame)
Call$required.vars <- c("elos", "k", "group", "regress")
Call$ncol.k <- 2
Call$ncol.elos <- 1
mf <- eval(Call, parent.frame())
if (nrow(mf) == 0)
stop("No (non-missing) observations")
Terms <- stats::terms(mf)
if (any(mf$adjust.A != 0))
warning("Any adjustments using 'adjust()' are being ignored.")
mf2 <- multiteam_model_frame(mf)
er <- elo.run(wins.A ~ elo.A + elo.B + k(k) + group(group) +
regress(regress, to = attr(mf$regress, "to"), by = attr(mf$regress,
"by")), data = mf2,
initial.elos = initial.elos,...)
er$Call <- Call
er$terms <- Terms
er$na.action <- stats::na.action(mf)
er$n.matches <- nrow(mf)
class(er) <- c("elo.run.multiteam", class(er))
er
}
Hi, first thanks for this great package, it is very interesting and nicely designed!
I ran into the issue with elo.run.multiteam where it ignores intial ratings so for now I rewrote the function to pass it to elo.run (not sure it s the correct way but looks ok)