> I really doubt the figure that only 15% of trades were done by algo a year ago. Maybe a few jumbo orders skew it, but by ticket numbers I would think over 80%

> I really doubt the figure that only 15% of trades were done by algo a year ago. Maybe a few jumbo orders skew it, but by ticket numbers I would think over 80% of trades are on some sort of automation. I'm inclined to agree with you. The article seems to conflate "algo" and electronic trading, which doesn't help matters. In my experience, the majority of FX trades are just straightforward deals at spot - no algo required. It's only when you've got a really big deal that you want to spread out to avoid moving the market, or where you need to execute at the fixing, that an algo would come into it. > "A year ago, algo orders placed by fund managers and other investors accounted for no more than 10-15% of total trading volumes.." To my mind "trading volumes" refers to the amount being traded, rather than the number of trades, so maybe he means that the value of the algo orders amounted to 10-15% of the total value of all trades (e.g. ten algo trades for $5m each vs one voice trade for $500m).