As you may be aware I have been carrying out some Premier League Predictions polls to start this season to see if the wisdom of crowds can predict the outcome of football matches, the hopeful aim is to carry out a poll for each match and see just how well social media and by extension twitter knows what they are talking about.
With two weeks of the season gone, I have carried out 20 polls and I can now give you a breakdown of what has happened
|Crystal Palace||Huddersfield||Palace fav||Incorrect|
|West Brom||Bournemouth||West Brom fav||Against||Incorrect|
|Brighton||Manchester City||Manchester City fav||Correct|
|Manchester United||West Ham||Man United fav||Correct|
|Swansea||Manchester United||Man United fav||Correct|
|Burnley||West Brom||Burnley fav||Against||Incorrect|
|Liverpool||Crystal Palace||Liverpool fav||Correct|
|Southampton||West Ham||Southampton fav||Correct|
|Manchester City||Everton||Man City fav||Incorrect|
Column 1 – Home Team
Column 2 – Away Team
Column 3 – Bookmakers Favourite
Column 4 – Whether The Prediction Went Against The Bookies
Column 5 – Outcome
Now if you are too lazy to count you can see that there have been 10 correct picks and 10 incorrect picks from the last two weeks Premier League action, a perfect 50/50 split in terms of outcome. What is interesting is that there have been 18 occasions in which twitter has gone with what the bookies reckon will happen and only twice where they bucked the trend.
What is interesting is that both games have involved West Brom. On the opening weekend they were favourites to beat Bournemouth, a result that came true but twitter thought that Eddie Howe’s men would win. Last week they were away to Burnley and twitter thought West Brom would win even though the bookies thought Sean Dyche’s men would come out on top.
So perhaps West Brom will be the wildcard in all of this, or maybe they have been in what may be considered close fixtures that could have gone either way. It has however shown that there is some favourite bias going on here as 90% of the picks have matched what the bookies have had down as favourites to win.
That may be just down to common sense though as if say Manchester United are playing West Ham at Old Trafford you are not going to get a large amount of votes for the Hammers to win, so the fact that twitter and the bookies are pretty much matching each other comes as no real surprise.
One thing that I have seen and again it is not something that has necessarily shocked me is the fact that Tottenham have been picked to win on both occasions. Both times they have been favourites to win but it is the sheer magnitude in which the votes have swung their way which has surprised me.
In the game against Newcastle, there were 0 votes for Newcastle to win at home. Again that may well be assumed but the fact that a large majority of my followers on twitter are Tottenham fans will have no doubt played a large part in that.
Time for some more "Scientific Polls" now
— Dan Tracey (@dantracey1983) August 12, 2017
The same can be said for the game against Chelsea last Sunday, they may well have been assumed favourites going into the clash with the Premier League champions but the manner in which the vote was split, is something of a surprise
— Dan Tracey (@dantracey1983) August 19, 2017
Here there were 82 votes and 71% of you thought that Tottenham would win, that is a much further spread than what the bookies believed – their odds before the game were
So Tottenham were the slight favourites but nowhere near the disparity in which we saw in terms of the votes. This would point to a large amount of club bias from the people who have taken the time to vote.
Another aspect I have been carrying out is putting my money where your mouth is, so I have been placing an accumulator based on your predictions (Saturday games only)
Just two picks out of 7 were correct, absolutely coupon busted
Slightly better with 4 out of 7. That means from the Saturday Premier League games in isolation Twitter has picked just 6 correct from 14 fixtures. They’ve only been £1 stakes though so no cause for concern, I don’t trust you that much!
The next step will be to carry out another round of picks this week and then we can start looking at correct home picks, correct away picks and correct picks for each club. Basically as the Premier League season continues we can take this all different directions. We’ve only just begun.
Please feel free to follow me on twitter and join in the polls and if this website has bought you there then please do say hello, also if you want to contact me in regards to any football data work that you currently have available then you can email me at firstname.lastname@example.org