'Considering the statistics'
by Black
2009-07-09 08:00
I did it, I considered the statistics.
When Trevor Manuel made that comment about considering the statistics, what he was implying was that PDV has achieved a 75% win rate (12 of 16) as national coach and therefore should be considered a success. This in statistical terms is known as a "hypothesis".
Now at face value this is a fair hypothesis, but seeing as I had the misfortune of doing statistics in my Masters year at UCT, I would like to challenge Mr Manual's hypothesis and as he suggested: in fact, "consider the statistics".
I remember it being drilled into me that when one seeks a true statistical result one needs to eliminate outliers and any other data which may skew the final result. In fact I remember being taught the following approach to "considering the statistics".
Step 1: Define the sample population - easy, PDV has coached sixteen test matches
Step 2: Eliminate outliers - an outlier would be like when the Boks beat the Blacks in Dunedin last year, it happens so rarely that it is not a true reflection of our ability to win. Another outlier was when we beat the Wallabies by almost 50 points; again such a large result is so rare it represents an invalid reflection of our ability to beat the Wallabies. So when "considering the statistics" we must remove both these two results from the sample population leaving us 14 remaining results to be considered.
Step 3: Bring normality to the sample population - basically what this means is that if we only ever played against countries we are expected to beat (eg Italy) we would have a 100% win rate thus giving an inaccurate result of our ability. Thus to provide a more accurate result of our win rate we must eliminate all the "sure things" as these skew the final result as well. In brief the seven games we played against Northern Hemisphere opponents and Argentina in 2008 must be eliminated as the opponents were so poor that beating them was not a true reflection of our ability to win. This now leaves us with only seven remaining results in our sample population.
Step 4: Calculate the true statistical rate from the remaining sample population - the seven remaining results consist of the four 2008 tri-nations matches we lost and the three B&I Lions matches of which we won two. Therefore by "considering the statistics" we can deduce that PDV has won only two of the seven matches that represent a true reflection of his ability thereby giving us a true statistical win rate of 28.6% Further more by "considering the statistics" we can conclude that Mr Manual's hypothesis (Pieter de Villiers is a successful coach) is flawed as PDV in true statistical terms only has a 28.6% win rate.
So there I did it, I considered the statistics and PDV is still poor!
Get published on MyNews24 by sending your letter, story or column to us.
Send us your news photos
Disclaimer: All articles and letters published on MyNews24 have been independently written by members of News24's community. The views of users published on News24 are therefore their own and do not necessarily represent the views of News24. News24 editors also reserve the right to edit or delete any and all comments received.