# Monthly Archives: June 2011

## Winsorization

Winsorization replaces extreme data values with less extreme values. But why Extreme values sometimes have a big effect on statistical operations.  That effect is not necessarily a good effect.  One approach to the problem is to change the statistical operation — this is the field of robust statistics. An alternative solution is to just change … Continue reading

Posted in R language, Statistics | | 9 Comments

## Review of “Adapt” by Tim Harford

The subtitle is “Why success always starts with failure”. Executive summary Brilliant.  Funny.  Enlightening. Complexity We live in a complex world. Exhibit number 1 (page 1) is the Toaster Project. Our brains are too small to know what to do in a complex world. The rules of adapting The three rules of adapting (page 36) … Continue reading

Posted in Book review | 2 Comments

## Talking The R Journal latest release

Volume 3/1 of The R Journal has been released. It of course has articles about using R.  In addition it has a feature that I highly support.  In preparation for the UseR! Conference 2011 August 16-18 in Coventry there are two Help Desk articles on making a good technical presentation. Technical presentation Rob Hyndman writes … Continue reading

## Highlights of the London Quant Group Technology Day

A summary of the high points of the day. Factor models and optimization Three of the talks formed a theme: factor models of variance — especially as applied to portfolio optimization. The basic problem is that variance matrices are created with error.  A variance matrix is a key input to (some) portfolio optimizations.  The optimizer … Continue reading

## Performance ratios, bootstrapping and infinite variances

If returns had infinite variance, would there be a problem bootstrapping information ratios? Background There is a discussion on the Quant Finance group of LinkedIn with the title: “How do you measure the confidence intervals of performance ratios?” One suggestion was to use the statistical bootstrap. This resulted in a discussion of the efficacy of … Continue reading

Posted in Performance, R language | | 3 Comments

## Bubble anatomy

Here is a schematic of a financial bubble. This is taken from a post by The Reformed Broker. Questions The picture feels right to me, but … Is there any data to support it? What process could fit a model like this without assuming the answer? Related posts Review of “Boombustology” by Vikram Mansharamani Subscribe … Continue reading

## Market arrows

Graphs like Figure 1 are reasonably common.  But they are not reasonable. Figure 1: A (log) price series with an explicit guide line. Some have the prices on a logarithmic scale, which is an improvement on the raw prices. The problem with this sort of plot is that two particular data points are taken as … Continue reading

## Premortem stress tests

To look in the corners we want to avoid. David Rowe’s latest Risk column, “Stress testing culture”, concerns the difficulty of developing good stress scenarios.  One of his particular concerns is that it is against our nature to consider the failure of something that we are working on.  We are masters at dodging that. There … Continue reading