Skip to Content

Continuing from my previous blog here. If you have not seen it please visit here to make sense of this blog.

http://scn.sap.com/community/lumira/blog/2013/09/27/lets-predict-olympics–rio-2016-part–1

Now let’s switch gears and find out the regression for one Olympic (let’s take Beijing). I will have to remove the China and Greece from the list because of HAF factor and Athens Olympics HAF factor. Also we will find out some outliers and see who they are and exclude from our analysis. Then probably we can speculate what might have caused this abnormal behavior.

/wp-content/uploads/2013/09/29_287210.jpg

Here are the outliers for the Beijing Olympic : Cuba , Hungary ,Jamaica,Japan,Kenya , Romania , Turkey.

Within these them Jamaica is behaving very strangely (May be we can term it a Bolt Factor 🙂 )

Here is how the data looks like /wp-content/uploads/2013/09/30_287211.jpg

Although these are the outliers ;  still looking at them looks like we can probably fit in a regression model (Although I wont try that adventure) in it and medal growth is  positively  correlates to GDP growth factor.

/wp-content/uploads/2013/09/27_287212.jpg

Now lets see how our average countries behave and they fit in real nicely. However there are two giants which are little anomalous as well (the two biggest countries in GDP, US and China; One growing very fast ahead of every one else in GDP growth and other far ahead of every one in absolute GDP).

/wp-content/uploads/2013/09/32_287216.jpg

And here is how every other measure stack up ; the trend is very clear. Either upwards or downwards.

/wp-content/uploads/2013/09/33_287217.jpg

Now lets fit in a regression model and start with a linear regression for these countries. We will start with a linear model and then we will take it from there/wp-content/uploads/2013/09/34_287219.jpg

Here is the result set after the run.  Although the goodness of fit is not great as countries like China and USA having defied the trend it 0.22

Here are the values that we care most.

The formula used is “Y = intercept + slope * X”
intercept -1.9209
slope 8.1467

The R-square factor is 0.2665

/wp-content/uploads/2013/09/35_287221.jpg

/wp-content/uploads/2013/09/36_287225.jpg

/wp-content/uploads/2013/09/37_287227.jpg

Now the final step is to interpret the results and try to predict the Rio games of 2016.

For that lets check the GDP growth rate of Brazil for last 5 years

2009  2010 2011      2012

-0.3        7.5          2.7          0.9

BRAZIL   Average GDP growth: 2.7 which is below average 

World Average : 3.02

GDP growth Factor : 0.894

So if the world maintains the same GDP growth and Brazil maintains the same Average GDP growth.

Y = intercept + slope * X”

The medal growth % of Brazil will be  = -1.9 + 8.1 * 0.894 = 5.54%
with intercept -1.9209 slope 8.1467

So what is the predicted medal tally of Brazil in 2016 Rio games ; but before that lets have a quick look how it has been performing in last 5 Olympic games.

/wp-content/uploads/2013/09/38_287231.jpg

The results are consistent if you discount the Sydney games although the size of observation is very low for any sort of conclusion.

Now here is the verdict based on our analysis.

Current Weighted medal tally = 28

Predicted Medal of Brazil in Rio :

( 28*1.76 (HAF excluding Athens))*1.05 = 51.72 ~ 52

If we make the medal shared in a similar ratio as in London 2012. Here is how it will

5 Gold, 11 Silver and  15 Bronze.

However we have to wait until  2016 to actually see how Brazil performs in Rio 2016.

Please let me know your thoughts and comments…

To report this post you need to login first.

2 Comments

You must be Logged on to comment or reply to a post.

Leave a Reply