After 10 rounds of the 2019 season I embarked on a wildly ambitious endeavour; to predict the results of every remaining match of the 2019 Season using a formula that weighed the current standings along with factors including home field, form and more.
Regardless of the result I promised to publish the results and for those of you who kindly took the time to read my article back on the 24/5/19, I had to follow through and review how my weekly predictions, projected ladder and projected Grand Finalists turn out. Here are the results of this crazy endeavour.
Below is the spreadsheet I used to predict the results (top half) and the actual results (bottom half). Green cells mean correct predictions and as the season went along the green becomes far more prevalent. Interestingly, it was a rough start with an under .500 success rate until round 14. From there on though it was much smoother sailing with a 69.51% success rate from round 15 through the end of the season. All in all this method of projecting did meet with some success, though it of course could not predict injury, rampant infighting (Knights), an incredibly unlikely and unsustainable and wildly improbable win streak (Panthers) and of course whatever that was (Dragons).
Final result: 62.83% Success rate in predicting the outcome of every game.
Expected ladder (based on the first 10 rounds)
Extrapolating out each teams win / loss to a full season shows the following Final Ladder.
Stats Projected ladder
Using the formula to project the results of each game and thus the ladder
|Club||Wins||Losses||Points||Rank Change from Expected Ladder|
|2.||Roosters||20||4||42||Up one spot|
|3.||Storm||19||5||40||Down one spot|
|4.||Sea Eagles||15||9||32||Up one spot|
|5.||Raiders||13||11||28||Down one spot|
|6.||Parramatta||13||11||28||Up two spots|
|7.||Knights||12||12||26||Down one spot|
|8.||Sharks||12||12||26||Up one spot|
|9.||Wests Tigers||10||14||22||Down two spots|
|10.||Broncos||9||15||20||Up one spot|
|11.||Warriors||9||15||20||Down one Spot|
|13.||Bulldogs||8||16||18||Up one spot|
|14.||Cowboys||8||16||18||Down one spot|
Actual end of season ladder
Final 2019 NRL Ladder
|Club||Wins||Losses||Points||Stat Project LDR Change||Expected LDR Change|
|1.||Storm||20||4||42||Up 2 spots||Up 1 spot|
|2.||Roosters||17||7||36||Unchanged||Up 1 spot|
|3.||Rabbitohs||16||8||34||Down 2 spots||Down 2 spots|
|4.||Raiders||15||9||32||Up 1 spot||Unchanged|
|5.||Eels||14||10||30||Up 1 spot||Up 3 spots|
|6.||Sea Eagles||14||10||30||Up 1 spot||Unchanged|
|7.||Sharks||12||12||26||Up 1 spot||Up 2 spots|
|8.||Broncos||11||12||25||Up 2 spots||Up 3 spots|
|9.||Wests Tigers||11||13||24||Unchanged||Down 2 spots|
|10.||Panthers||11||13||24||Up 6 spots||Up 6 spots|
|11.||Knights||10||14||22||Down 4 spots||Down 5 spots|
|12.||Bulldogs||10||14||20||Up 1 spot||Up 2 spots|
|13.||Warriors||9||14||21||Down 2 spots||Down 3 spots|
|14.||Cowboys||9||15||20||Unchanged||Down 1 spot|
|15.||Dragons||8||16||18||Down 3 spots||Down 3 spots|
|16.||Titans||4||120||10||Down 1 spot||Down 1 spot|
- Note: There was one draw in 2019 (Warriors vs. Broncos)
Average Ladder Position Variance
|Stat Projected Ladder||Expected Ladder|
Success!!!! The stat projected ladder was on average out by 1.68 spots on the ladder for each team while the expected ladder was out by 2.06 positions! The stat projected ladder was in fact a more accurate means of predicting the final ladder than merely extrapolating the Rd. 10 ladder. Was it worth all the work? Debatable. Was it as accurate as I would have liked? Of course not, but that is half the fun and next years installment will feature tweaks to the formula to try to improve on the 2019 Variance.
Final result: The stats projected Ladder was more accurate in predicting each teams final ladder position.
|Category||Stat Projected Ladder||Expected Ladder|
Dead even with the correct predictions using the two methods. While the per ladder position result using the stat projected was better than the predicted ladder, each was just as effective at accurately predicting a specific result. Interesting.
The projected Roosters v Raiders Grand Final is still alive and for those who saw their week one finals efforts, looks a very real possibility. We shall check back on this in 2 weeks time.
Final Result: To be determined
show filter2 big movers in Penrith and Newcastle defied their projections for very contrasting reasons. While the Panthers got hot and went on a completely unexpected winning streak, the wheels completely fell off for the Novacastrians, resulting in the highly contentious departure of their coach. We know we can’t use the stats to predict everything and these two teams absolutely exemplified this in 2019.
Penrith’s win streak came from out of nowhere after a disastrous start to the season while the alleged infighting and back room moves at the Knights can be largely attributed to their downfall, but we cannot quantify it. While we typically know a coach losing the team will have a negative impact on the side, this exercise has shown there is no measure for it nor any way to predict it.
The stats can give us tremendous insight into our great game, but at the of the day there will always be the human element to it, and as this exercise has demonstrated we cannot put a number on that…… yet.
Thanks for reading as always and good luck to those with teams still in the hunt.
Trust in the Data!