People from 35 countries around the globe have chosen a winner!
To be exact, 126 people, about two thirds of whom live outside of the UK, have chosen (*). Thank you, everyone who completed the survey. The number of respondents really took me by surprise because I expected to get only ten to fifteen, at most twenty, responses. The participants helped to predict the results of the 17 remaining matches (at survey time) for each of the current top three teams: Chelsea, Man City, and our Arsenal. On debate is the question, “Will Arsenal remain ToTL at the end of the season?”
Yes, We Can
Without further delays lest I incur the wrath of evonne, the survey says: Chelsea third with 85 pts, City second with 87 pts, and Arsenal comfortably Top of The League with a two-win gap at 93 pts! Surprised?
Plotting out the predictions and point totals each week, the graph below shows how people see the rest of the season to play out for the three teams.
Perhaps the final position predictions are not a surprise to AA readers, but there are some interesting match forecasts hidden beneath the overall totals.
Chelsea vs Newcastle: I had this down as a sure win for Chelsea, but almost half of you predict otherwise, with 40% of the people seeing a draw. A possible two point drop for Chelsea here, although that alone would not affect the final three-team ranking.
Spurs vs City and ManU vs City: Can Arsenal’s neighbors help them out next week? Almost 50% of the people expect Spurs will get a point from this fixture. More importantly for Arsenal, that would mean two points dropped by City. Similarly later on in the season, City face an away derby with Arsenal other historical rival, Man Utd. Would we wish BSR and that gang success on the day? 50% of the people pine for a draw out of this match as well.
Arsenal vs City and Spurs vs Arsenal: In the City home fixture, almost 60% of the people predicted a win for Arsenal, with almost 40% thinking draw. Only a small percentage were pessimistic. However, if it turns out to be a draw instead, suddenly the point prediction would become Arsenal 91 pts, City 88 pts, and that final fixture becomes important again. The percentages and prediction in the Spur away match are similar. Again, a draw instead would see Arsenal dropping to 91 pts, but in this case there is no gain for City.
Chelsea vs Arsenal: Close to 50% of the people see this as a draw. 35% predict a win for Arsenal, and 15% predict a loss. The uncertainly is high for this match.
Arsenal to go undefeated the rest of the way? People seem to think so. They predict clear wins for many matches. The most uncertain fixture in the minds of the respondents is Chelsea away, but even for this, only about 15% of the people expect a loss. In the other high profile matches, 6% see a loss to Liverpool, 4% to City, 4% to Everton, and 2% to Spur and Man Utd. The rest are all 1% or less!
How does the survey compare to other opinions out there? After all, in predicting a draw for Chelsea vs Man Utd, it already got one wrong.
Experts Disagree
The well known pundits have not been shy to opine about Arsenal. And how can they avoid it, really? Arsenal’s performances so far have forced them to speak up.
- Expert 1, Alan Hansen 27 Oct 2013: “[Arsenal] will not end the season as champions.”
- Expert 2, Michael Owen 18 Dec 2013: “Arsenal don’t have the class of player to go toe to toe with the main title contenders.”
- Expert 3, Gary Neville 14 Jan 2014: “I do think Arsenal will get hauled back by those two teams.” He meant Chelsea and Man City of course.
- Expert 4, Robbie Savage 14 Jan 2014: “But will they [Arsenal] win the title? I still say no.”
- Expert 5, Alan Shearer 18 Jan 2014: “I didn’t think they will [win the title], and I don’t see any reason why I should change my mind.”
And so on… You get the idea. Although recently the pundits have begun—some of them begrudgingly—to compliment Arsenal’s good performances to date, they are still firm in their belief: Arsenal ain’t gonna to make it.
Confidence or Bias?
A few years ago, a book titled The Wisdom of Crowds, by James Surowiecki, came out. In it, the author shows that, in certain situations, an unbiased sampling of the collective judgement can be very accurate, even more accurate than that of experts. Will it happen here? Now, it is hard to claim that readers of AA, although level headed for the most part, are unbiased when it comes to Arsenal (*). In fact, I was quite surprised at the level of confidence the respondents have in the Arsenal.
For me, this is a significant, and unexpected, revelation of the survey, that this global group of fans believe so strongly in the Arsenal. In three quarters of the remaining Arsenal fixtures, about 1% the predictions foresee a loss. In only two games are the loss predictions more than 5%, for a measly overall loss percentage of 2%. Confidence or bias? To contrast, this same group of people sees the loss percentages of Chelsea and City to be about 10%, neither of whom is predicted to go undefeated the rest of the way (which, as others have pointed out on AA, feel more realistic).
Furthermore, has this confidence/bias always been so strong? Unfortunately I don’t have data for other years, but judging by the various views expressed on AA and elsewhere, I doubt it. It feels like emergent confidence when 80% of the people expect a win in the upcoming home game against Man United, and only 2% seeing a loss. People seem more comfortable that the Arsenal this year can overcome the challenges that tripped them up in the last few years. From the survey, I cannot detect the doom-and-gloom mood amongst the fans, which appeared even as recently as the opening match this season. One can of course still hear or read the public grumble and negative comments, but this survey indicates that a strong counterview exists.
We have to wait to the end of the season to know who got it right, the fans or the experts. Until then, keep believing; it can happen again this year.
Move On! Nothing More To See Here…
Norfolk, look away! For the rest, who like numbers, tables, and charts, have fun with the information below.
These columns show the game-by-game predictions of win/draw/lose by percentage of respondents.
The next set of columns are the game-by-game point averages and the corresponding predictions. The circles are the translation of the survey tallies into W/D/L (†). The error bars are the standard deviation of any individual prediction from the point average of the corresponding match (‡).
If the above doesn’t provide you enough to analyze, pls write me via AA, and I can give you more info.
Notes
(*) There were actually a few more than 126 participants; however these additional people were clearly over-enthusiastic—they selected all wins for Arsenal and all losses for the other two teams—that I had to drop their responses.
(†) For each match, we have the tallies of win, draw, and lose choices. If the win tally is greater than the sum of draws and losses, then this is considered a win. Likewise, if the loss tally is greater than the sum of draw and win, then it is a loss. Otherwise it is a draw. For example, the tallies of 51 W, 40D, 9 L would translate to a win, while the tallies of 25 W, 35 D, 40 L would map into a draw. This mapping seems to serve the purpose in this instance, but one can imagine it not working if the opinions were polarized.
(‡) Point average is done by weighing the points from a win, with the percent of people predicting a win, and so on. For example, if 55% of the people expect a win, 40% a draw, and 5% a loss, then the point average would be (0.55 * 3 + 0.40 * 1 + 0.10 * 0) = 1.93. If we have 25% W, 35% D, and 40% L, this would give a point average of 1.1.
Written by TT