The Numerical Advantage: 2020 Finals Predictions
Starting Round 11 last season, this column predicted the winner in 62.83% of games using the data from the first 10 rounds, including predicting both Grand Finalists and the Premiers! How will we go in 2020 using only last season’s data to guide us…
Last year’s experiment in predicting the rest of the season's outcomes using the data from the first 10 rounds certainly exceeded my expectations. If I was a smart man I would probably walk off into the sunset and rest on those achievements for remainder of my days.
But I’m just not that smart. I can’t help but wonder... can we do it for the whole season? Can we predict the Grand Finalists before the first Kick Off? Can we crack the 70% accuracy mark… surely with just a few tweaks to the formula…. Damn it, I have to know! So here we go again ladies and gentlemen, your data driven predictions for the 2020 NRL season.
Factors
Last year’s predictions used a fairly rudimentary calculation to score each team for each game with the team with the higher score predicted as the winner. This was based on ladder position, opponent ladder position & home field advantage. This was successful 62.83% of the time, but this year we are going for 70% accuracy and we don’t have the benefit of 10 rounds of form to make our predictions. To create a more comprehensive system we need to consider what really matters in determining a winner.
Is playing at home still an advantage?
Home field is no longer the great advantage it once was. Back in 2000, home teams won a whopping 65% of games, basically two out of every three. In recent times with teams sharing venues, playing at neutral venues for cash and the improved travelling conditions as the game has become yet more professional, this rate dropped as low as 53% in 2015, improving to 56.25% in 2019 – still well down from the turn of the century success rates. Over the last 5 years it is 55.27%. Nine teams in the past 5 years have missed the finals despite finishing with a winning home record.
One of the key learnings from our last attempt is that home team success, 67.44% after 10 rounds, came back to the pack over the course of the season 56.25%. Whereas last year home field was worth 4 point in the calculations, this years will be scored with 3 points for a home game or a 1.5 point score for the listed home team at a neutral venue.
Note that the Roosters games all around Sydney are not included as neutral in this as there is minimal travel for them so they are not disrupted in the manner they will be when travelling to Perth in rd. 11 for their ‘home’ game against the Raiders. Sharks home games are at Jubilee.
Significant home games at Neutral venues |
Roosters (H) v Manly (A) - Central Cost Stadium, Central Coast |
Raiders (H) v Knights (A) - McDonalds Park, Wagga Wagga |
Eels (H) v Broncos (A) - TIO Stadium, Darwin |
Tigers (H) V Raiders (A) - Scully Park, Tamworth |
Panthers (H) v Sharks (A) - Carrington Park, Bathurst |
Sharks (H) v Cowboys (A) - Sunshine Coast Stadium, Sunshine Coast |
Sea Eagles (H) v Tigers (A) - Central Coast Stadium, Central Coast |
Roosters (H) v Raiders (A) - HBF Park, Perth |
Rabbitohs (H) v Raiders (A) - Central Coast Stadium, Central Coast |
Magic Round
Once again the Broncos have somehow scored an away game at SunCorp for Magic Round, while Wests, Manly, Souths, Canberra, Roosters, Warriors, Melbourne & the Titans have given up a home game. While this weekend is terrific for the promotion of the game, it will result in a deduction off the home score for each of these faux ‘home’ teams.
The effect of short turnarounds
Was there a correlation between the amount of days between games and the results last year? The following table presents the winning % for each team based on their turnaround from their previous game in 2019.
Days | Winning % |
5 | 36.00 |
6 | 51.30 |
7 | 53.01 |
8 | 50.70 |
9 | 39.29 |
10 | 41.67 |
11 | 0.00 |
12 | 60.00 |
13 plus | 52.38 |
*note: there was only one 11 day turnaround, which was a loss.
In 2019 Short turnarounds (5 days) did have a negative effect (36.00%), while interestingly 9 & 10 day turnarounds were also less successful (39.29% & 41.67%). 6, 7, & 8 day turnarounds were more successful with 12 or more days being the ideal. Taking this into account there will be a negative score attributed to teams coming off a 5, 9 or 10 day turnaround, a positive score for 6,7,8 & 13 days, and a higher score for teams coming off a 12 day layoff.
Days | Points |
5 | -2 |
6 | +1 |
7 | +1.5 |
8 | +1 |
9 | -1 |
10 | -1 |
11 | 0 |
12 | +2 |
13 plus | +1.5 |
The Formula
ladder position, +/- home field advantage, +/- turnaround
minus opponent’s ladder position, +/- home field advantage, +/- turnaround
The team with the higher score in each matchup will be our predictive winner.
Predicted Ladder
After applying our formula to every single game for the upcoming season, here is the projection.
Position | Team | Win | Loss |
1st | Melbourne Storm | 22 | 2 |
2nd | South Sydney Rabbitohs | 21 | 3 |
3rd | Sydney Roosters | 20 | 4 |
4th | Canberra Raiders | 19 | 5 |
5th | Parramatta Eels | 17 | 7 |
6th | Manly Sea Eagles | 15 | 9 |
7th | Cronulla Sharks | 14 | 10 |
8th | Brisbane Broncos | 12 | 12 |
9th | Penrith Panthers | 10 | 14 |
10th | Wests Tigers | 10 | 14 |
11th | Canterbury Bulldogs | 9 | 15 |
12th | Newcastle Knights | 9 | 15 |
13th | New Zealand Warriors | 4 | 20 |
14th | St George Dragons | 4 | 20 |
15th | Gold Coast Titans | 3 | 21 |
16th | North Queensland Cowboys | 3 | 21 |
Now obviously this a projection. This does not take into account injuries, suspensions, coaching changes, etc. This is simply an attempt to develop a formula to predict the season before it kicks off based on last years results, so please keep this in mind when hitting up the comments section. There is not a massive amount of change to the ladder positions from 2019 (poor Cowboys) but no team is projected to finish with their same record as 2019, which was surprising.
Predicted Grand Finalists
Using the same successful formula as last year, where we had the Roosters vs the Raiders as our GF, we know that the Grand Finalists are most likely to be the teams that finish 2nd and 5th. Using this as our guide again our Grand Finalists for 2020 will be the….
Rabbitohs and Eels
Predicted Premiers
With our eventual premiers being… The South Sydney Rabbitohs
So there it is folks. There will be 192 games in the 2020 NRL Regular Season. To get to our 70% benchmark we need to hit on 135 of them. Wish me luck. This should be interesting.
As always, thanks for reading. Comments & thoughts are always welcome in the comments section and best of luck to your club for the upcoming season.
Trust in the Data!