Predicting stock market doom using networks analysis

Writing in the Financial Times, Bob Wigley, Chairman of UK Finance, describes some joint work we are doing to understand structural changes to markets in the run-up to crashes and other discontinuities. The basic idea is that, while obvious indicators like the VIXX or the performance of a stock market index only change once something bad is underway (and it's too late to do anything), subtler, meso-scale measures often presentiently undergo substantial deviations from normality long (months) before the crash. 

Market Pressure Gauge Flashes a Red Alert

I was talking to a friend about this who said, 'it must be exciting, but it also seems very lonely, to be the predictor of doom'. Yeah, that's about right.