allowayai

AI and Sports Analytics Utility Toolkit for R

allowayai is an R package built around the workflows I write about on AllowayAI — sports analytics, machine learning model evaluation, and practical data science utilities. It is designed to be a clean, well-documented toolkit that bridges the gap between statistical theory and real-world betting and prediction workflows.

Installation

You can install the development version from GitHub:

# install.packages("devtools")
devtools::install_github("ianalloway/allowayai")

Functions

The package contains the following unique functions:

Function	Type	Description
calc_kelly()	S3	Calculates optimal bet sizing using the Kelly Criterion
print.kelly_bet()	S3 method	Pretty-prints a kelly_bet object
make_sports_model()	S4 constructor	Creates a SportsModel S4 object
show() for SportsModel	S4 method	Displays model performance summary
summarize_types()	Utility	Returns a data frame of column names and types

Usage

Kelly Criterion Bet Sizing

library(allowayai)

# Calculate optimal bet size for a 60% win probability at 2.0 decimal odds
bet <- calc_kelly(prob = 0.60, odds = 2.0, fraction = 0.5)
print(bet)
#> --- Kelly Criterion Bet Sizing ---
#> Model Probability: 60.0%
#> Implied Probability: 50.0%
#> Edge: 10.0%
#> Full Kelly: 20.00% of bankroll
#> Recommended Bet (0.50x): 10.00% of bankroll

S4 Sports Model Object

# Create a sports model performance record
model <- make_sports_model(accuracy = 0.68, roi = 0.12, games_tested = 6000)
show(model)
#> Sports Betting Model Performance
#> --------------------------------
#> Accuracy: 68.0%
#> ROI: +12.0%
#> Sample Size: 6000 games

Data Frame Type Summary

df <- data.frame(
  team = c("Chiefs", "Eagles"),
  wins = c(14, 13),
  playoff = c(TRUE, TRUE)
)

summarize_types(df)
#>    column      type
#> 1    team character
#> 2    wins   integer
#> 3 playoff   logical

Package Structure

• S3 classes and methods: kelly_bet class with print method
• S4 classes and methods: SportsModel class with show method
• Utility functions: summarize_types()
• Vignette: Full walkthrough in vignettes/allowayai.Rmd
• Tests: testthat unit tests covering all functions
• License: MIT

Author

Ian Alloway — ianalloway.xyz | AllowayAI on Substack