And the likely answer is that with enough detailed analytical work on the available social media data that we could get down to a small group of prospective winners for each year’s race.
For example, this year with the race literally just finished, a quick analysis of the social data streams shows that over the PRIOR 3 days there were about 8,000 tweets which referenced themselves as located in Australia and referring to the favourite horses.
Associated with that Twitter data is a wordcloud, which shows the most frequently occurring words graded by font size. The larger the font the more often the word occurred, and the results below are clear.
Our analysis took just ten minutes using Sysomos, so with more detailed and painstaking analysis a more precise result could be achieved.
Twitter ranks differently to betting
When we review the Twitter data in the wordcloud above it shows that the most frequent names in the 3 days prior to race day were, in rank order:
- Ethiopa, followed by
- Green Moon, and then equal
- Winchester and Dunaden, and then
According to the bookmakers near the day of the race the ranking was this (Sportingbet Australia Market):
- Americain $6.00
- Dunaden 6.50
- Mount Athos 9.00
- Red Cadeaux 11.00
- Brigantin 13.00
- Maluckyday 14.00
- Ethiopia 14.00
- Shahwardi 21.00
- My Quest For Peace 21.00
- Green Moon 21.00
- Lights Of Heaven 23.00
- Tac De Boistron 23.00
- Gatewood 26.00
- Fiorente 26.00
- Galileos Choice 26.00
- Winchester 31.00
Which is a very different ranking to the social media. In the press and in the minds of the betting businesses Americain was a clear favourite and Green Moon, Ethiopia, and Winchester were well down. In Twitter Americain was ranked 5th behind those other two.
Green Moon stuns stellar Cup field: GREEN Moon has won the $6.2m Melbourne Cup with Brett Prebble on board in a … bit.ly/YzCrgR
— The Australian (@australian) November 6, 2012
The first three across the line were Green Moon ($21) defeating the Gai Waterhouse-trained Fiorente ($31) by a length, with Jakkalberry ($81) a length and a quarter away third.
Comparing different media we get different views
Comparing those 5 top twitter horse name across different media becomes interesting.
Across all media, social and online traditional, Americain tops out.
Things become even more interesting when we compare between media, where Americain dominates News and Blogs, Ethiopia tops Forums, and Green Moon tops Twitter. Remember these are all in the 3 days before race day. So if you were a social media geek and a sporting person you might just place some bets on Ethiopia and Green Moon, based on these results.
After the fact – no better prediction for place getters
If we compare the actual race place-getters against the Twitter wordcloud favourites across different media then we do not see signs that those place-getters, other than Green Moon, would have been highlighted by any media analysis. Both Fiorente and Jakkalberry are off the radar. That’s not so say that by continuing to work backwards and refining the approach we may not be able to bring them to the fore, but in terms of this quick analysis we would have missed them – as most punters did on the day.
A quick analysis of social media data gave a clue that Green Moon was an under-backed contender for the Melbourne Cup, but on face value this analysis did not predict the likely place-getters. With more specific and detailed analysis we think it is likely that winners and place-getters could be relatively accurately predicted, taking into account chatter right up to the hour of the race for example.
Tweet to @igo2Follow iGo2 on Linkedin
Search for: Americain OR Dunaden OR “Mount Athos” OR “Red Cadeaux” OR Brigantin OR Maluckyday OR Ethiopia OR Shahwardi OR “My Quest For Peace” OR “Green Moon” OR “Lights Of Heaven” OR “Tac De Boistron” OR Gatewood OR Fiorente OR “Galileos Choice” OR Winchester OR Mourayan OR “Glencadam Gold” OR Jakkalberry
Demographics: For Wordcloud restrict to Twitter located in Australia; Comparisons are all media.
Eliminate noise words: today; run; justhorseracing; melb; think; racing; racinginsider; melbcup; great; 2012; melbourne; melbournecup; cup; horse; plate; horseracing; caulfield;
Postcript – data from News not of value
Interestingly while the Twitter wordcloud enabled us to deduce the potential outcomes, the wordcloud generated from News gave no clues. It was not for lack of data since online News was the second most popular source of mentions about the Cup and the horses. As can be seen below, even if most of the filler words are eliminated there is little to go on in terms of favourites. Perhaps it’s a hidden benefit of tweets that the 140 character limit keeps things on the topic, and the News tends to fill the space and also try to fill stories with gossip and such.