{"id":8990,"date":"2016-10-03T05:14:36","date_gmt":"2016-10-03T05:14:36","guid":{"rendered":"https:\/\/www.techdesignforums.com\/practice\/?p=8990"},"modified":"2016-10-03T05:17:37","modified_gmt":"2016-10-03T05:17:37","slug":"wally-rhines-picks-the-signal-from-the-noise","status":"publish","type":"post","link":"https:\/\/www.techdesignforums.com\/practice\/technique\/wally-rhines-picks-the-signal-from-the-noise\/","title":{"rendered":"Wally Rhines separates the signal from the noise"},"content":{"rendered":"

A cynic will say there are only two types of forecast: wrong and lucky. That isn\u2019t true, the sports-to-politics statistician Nate Silver has an uncanny ability to call events. He has written one of the best explanations as to why forecasts go astray, The Signal and The Noise<\/a> \u2013 if you are a forecasting or math buff, it\u2019s worth your time.<\/p>\n

But Silver has not applied his skills to silicon. Someone who has is Wally Rhines, chairman and CEO of Mentor Graphics<\/a>. Rhines has a reputation in the business for \u2018doing the math\u2019 and doing it well. He is sharing his latest observations in a series of global keynotes.<\/p>\n

Rhines\u2019 latest observations come in two parts.<\/p>\n

First, he has dug into the reasons why semiconductor growth and revenue forecasts so often miss the mark. Last year analysts\u2019 expectations for 2016 coalesced around 3.5% growth. \u201cBut one year later, and two-thirds of the way through 2016, the numbers are down by 0.8%. That\u2019s a big swing,\u201d he notes. Rhines has some solid ideas for what needs rethinking.<\/p>\n

Second, Rhines is proposing a way of detecting future sources of growth \u2013 particularly in the Internet of Things – that could better indicate both where the market is going overall and what specific segments most justify R&D investment.<\/p>\n

To do that, he has applied a well-known technique from fields as different as cancer rate forecasts and software project management: The Gompertz Distribution.<\/p>\n

Silicon\u2019s \u2018hidden\u2019 sales<\/strong><\/h3>\n

Wally Rhines argues that a major challenge when forecasting semiconductor growth is that you must account for two types of recession: the macroeconomic and the silicon-specific when supply shoots past demand. It does not help forecasters that the two types do not occur simultaneously. \u201cThese variables don\u2019t change together,\u201d noted Rhines.<\/p>\n

He starts his analysis by highlighting some stable market data.<\/p>\n