John Hooper Medal for Statistics Competition 2020

Luke & Liam used secondary data obtained from Cinebench's publicly available online score catalogue and manufacturing listing. Following an analysis of the relationship between power and performance of consumer CPUs their results:

·         confirmed their hypothesis that CPU power consumption and CPU performance are still closely linked, as power use and “clock speed” are linearly proportional to each other.

·         identified that the strength of the correlations observed, especially between TDP and CB multi-core score (R2= 0.487) make it clear that there is much improvement to be made in the sector. As this difference is due to energy used in the overhead running of the chip, they identified this as the area with the highest potential for efficiency improvements.

In conclusion, they feel that a higher emphasis on power to performance ratio is crucial to CPU design if future energy use within the technology industry is to be in any way sustainable.


We would also to take this opportunity to thank our other very talented first year teams who entered the Competition. Alina Siddiqui, Zofia Bielicka and Zuzanna Giermaz conducted the following statistical analysis: “Are stereotypical variables about being an Ambivert/ Extrovert/ Introvert valid?” and Aisling Harte and Isabelle Butler-Clyne produced a statistical report: “First Year Phone Usage Survey Project”.

Other News