We've found alot of their articles great reads on the pulse of innovation and rebalancing of informational power between the buyside and sellside. Unfortunately sometimes the really great stories don't make it to their website - like the one we read in the July-September 2005 issue titled, "Sketches on the algorithmic canvas". The article is part of a series based on an ongoing survey of buy-side attitudes to algorithmic trading across the US, UK and continental Europe. Firms like Morgan Stanley, Lehmans, CSFB, and Goldmans are evaluated by the head traders at client firms. The articles also discuss how smaller client firms use the algorithm tools these big companies provide. The results are extremely interesting. There are varying degrees of enthusiasm for these tools - with a lot of hype that confuses what these tools really provide in terms of cost benefits or return ("alpha") benefits.
I like the thoughts expressed by Simon Thompson, head of equity portfolio trading for BGI and an active user of algorithms.
"What algorithms add is expedience. If you are relying on algorithms to provide you with alpha, then your investment management model is lacking some robustness."Makes sense: these tools help deliver on the cost of trading/working an order - they are not going to haul in the outstanding returns.
Here's the meat of the article that really got my attention: The article then quotes Steve Wood, global head of trading at Schroders.
'Sniffer' programs like Trade-Ideas indeed. There are already several firms using Trade-Ideas to act as their sniffer against any undue market impact when running their various algorithms. If the impact of their activity appears within Trade-Ideas, they switch to the human touch. Trade-Ideas acts as a first line of defense - their tripwire. Conversely we have many individual subscribers who hunt for this sort of activity and prey on those who don't employ Algorithmic Monitoring tools.
"We use a mix of DMA (direct market access tools (i.e., manned by traders)) and algorithms. We focus on putting the vast majority down an algorithmic pipe, but then looking at the order flow that is moving away from us or moving for us, taking it out of the algorithmic trading mechanism and using DMA to add value. So we get a hybrid of both, which enables us to reduce our costs.
That approach also makes it less likely that the strategy will be detected by a 'sniffer' program."
Our conclusion: It's a good position to be in when both sides of the contest find your tool invaluable.