Current AFL Mens ratings and predictions

Welcome to our all new AFL Men's ratings and tips page! Here you will the latest ratings, tips for the upcoming rounds and any simulations that have been done. I'll usually aim to update these early in the week to give you time to digets. Any questions, let me know on Twitter!

Read more: How our ELO rating system works | Simulating the season | A review of 2016 ELO performance

Current ratings

These are the current AFL Men's ratings. The horizontal lines represent a the top 4 and top 8 cutoff's respectively. During the season, we'll also display any changes to ratings on the graph and the movement of teams up or down the ladder. If you want to know how these are calculated, you can read our methodolgy here.

plot of chunk ratings_plot

Tips

Each week our tops for the upcoming round of matches will be here. Currently, our tips for 26 in in! See the predicted margin and probabilities for each below. I'm playing around with using some packages to make a nice table but for now, I'll keep uploading the Excel image. You can view access the R Markdown table below.

plot of chunk unnamed-chunk-1

Simulations

During the season, each week we run a simulation of the remainder of the season. You can read the full methodology here. In short, we take the current ELO ratings and use those to set an expected margin for the upcoming set of games. We then randomly draw a margin using that expected margin as our mean, and the historical standard deviation in margins as our error. From this, we get a 'simulated' result. We update the ELO rating, move onto the next round and keep going until the season is over. We then repeat this 10 000 times to obtain our probability estimates of each finishing position. They are listed below!

Firstly, we have a heat map showing the expected finishing position of each team. Darker red indicates a higher probability (with the actual percentages appearing in the boxes). I've also added cutoffs for top 4 and top 8.

Next we have our win distribution for each team. These will get tighter as the season progresses but in summary, they show the distribution of expected wins in all of our simulations. The vertical line is the average number of wins.

Finally, we have our simulation table below. This shows us the probability of finishing in various important positions on the ladder. As we get towards the end of the year, we can mark off when teams can or can't make certain spots.

Tipping table

This is here purely for testing purposes. I'm playing around with making a nicely formatted table in R Markdown using Formattable. This is the result so far – pretty ugly but I'll play around with it!

Round Game Game_all Date Home Home.ELO Away Away.ELO Margin Probability Result Label
26 1 15198 Fri, Sep 22 Adelaide 1632 Geelong 1599 16 59.66% Adelaide by 16 Adelaide v Geelong
26 2 15199 Sat, Sep 23 Richmond 1581 GWS 1591 6 53.59% Richmond by 6 Richmond v GWS