Third and last episode of "The Code"

Interesting final programme which covered a lot of ground but seemed slower than the other two episodes. It focused upon mathematical prediction and started with Christopher Columbus gaining much needed supplies from unfriendly (and superstitious) natives by using tables that predicted eclipses.
Next Marcus avoided a 30Kg steel "ball of death" by predicting it's trajectory and landing point. He then pointed that the real world seems more random and less well-defined. Which moved us onto the flight behaviour of swarms of migrating starlings in Denmark which were able to be mathematically modelled using only a few parameters (For more info on how this complex behaviour can be modelled with simple rules search on "Agent based Models". I can recommend a free program called Netlogo and it's models of swarms).
This kind of software can be used to model all sorts of scenarios including competition within bacterial populations or (as a an example of a complex model) contamination events on food-production lines.
The next question was could this be extended to human behaviour and we are shown that it can be e.g. in modelling evacuation of buildings to design safer and more efficient ones).
The point was made that humans tend to behave in patterns with examples of the Rock, Paper, Scissors game and how detectives can use this kind of mathematics to track serial killers (This formed the basis of the pilot episode of the hit TV-drama series Numb3rs). Games like RPS can lead you to examples like non-transitive dice games which can also be described by the mathematical areas of Group Theory and Game Theory (see http://www.ansible.co.uk/misc/mgardner.html).

Marcus then debunked the urban legend of Lemmings mass suicides and showed using a simple equation that described population growth with the two key properties of current population and growth rate. Lemmings have an unusually high population growth rate that causes dramatic "boom-bust" behaviour which is chaotic rather than random. Again, if you are interested in this area (mathematical modelling of populations, predator-prey etc)  I can recommend another piece of free software called Populus.
Chaos then led us to a famous British obsession - the weather. M du S quickly showed how chaotic behaviour limited weather prediction to short-term effects.

We then switched to the "Wisdom of the Crowds" and briefly mentioned Galton's Ox before moving on to the more popularist "How many jelly-beans in the jar" problem (See an extremely good example here on YouTube featuring Professor David Speigelhalter). The take home message was that when large numbers of opinions were solicited, then the over-estimators compensated for the under-estimators by cancelling out the errors. This wisdom of the crowds lead us to Google who explored using people's searches and relating them to real world trends such as forthcoming influenza epidemics. Twitter mining was also mentioned relating negative words amongst traders to the behaviour of the stock market.

The episode (and series) ended on an almost-incongruous idea that related  the doubling of a cities size resulting in a 15% increase in other socio-economic factors (e.g. salaries). The idea matched well with Andrew Marrs series on Mega-cities but I found it to be a weak ending to an otherwise excellent math & science communication programme.