A good starting point is...

A good starting point is here:

http://en.wikipedia.org/wiki/Algorithmic_trading

I guess 'bot' is short for 'robot', rather than 'bottom' or 'botanical'.

In summary, bots are black box computer programs which execute trading strategies.

For proprietary traders (such as CLSA, Tibra and many others including investment banks) this means either buy/sell or sell/buy over time (intra-day or across many days) at a profit.

For brokers, algo trading is used to work large client orders. So if you are an insto client and place a huge order with your broker, they may choose to use a bot which will work the order over time rather than go to market with the whole order at once. We saw this when Orbis took their rather large stake in CER through Deutsche Bank. There simply weren't enough shares in-market, so DB worked the order over a period of many weeks assisted by a bot. Those of you who lived through this will know how effective this strategy was.

A number of long-term strategies in lower liquidity stocks (i.e. those which are able to be influenced - not your BHPs of the world) involve bending the technical indicators in your favour. I made a posting recently about a Keynesian Beauty Contest which can be used to describe this. It takes a while to get your head around it.

There are many, many strategies, but here is an example of one:

You are an investment banker (your parents are so proud) and want to buy a lot of stock X for a client... way more than is available in market. If you understand the rules of technical analysis, you will know that if you can get the price action over time to signal a very bearish pattern (intra or inter day), traders are likely to sell. In time, the sell action becomes a self-fulfilling prophecy as downward price pressure momentum builds (that little voice in your head which says "oh dear the price has gone down, somebody knows something, I should sell too").

To get the best deal for your client (who is a buyer of the stock), you initially need a bot strategy which causes the price to exhibit bearish patterns, such as giving the appearance that a lot of selling is going one. In this example, sell side stacking is used as an algorithmic lubricant in the process. Then you need a strategy to fulfill the client order over time as the opportunities present.

The guys and girls who build the bots are well educated and well paid. Here is an example role with Optiver:

http://executive.seek.com.au/users/apply/index.ascx?JobID=16159823

...and here is the company web site:

http://www.optiver.nl/the-company/locations/optiver-asia-pacific/

Moving back to CER... we have no idea what strategy was being played out today and probably never will. I personally find that a little knowledge and the ability to recognise the activity helps me understand the bigger picture. Hopefully you feel the same way too.