For most veteran traders, and many technologically savvy traders too, algo trading is a daunting challenge to face. This is especially true if you are a micro trend trader, a scalper or a screen trader. In any case, if your trading regimen requires speed of execution and order fills, algo traders can and do upset your apple cart/s. While I dont want to build a case for and / or against the merits and demerits of algo trading, I as a trader, a founder promoter of an incorporated entity that depends on trading for bread & butter, am surely concerned about getting around the challenges posed to my trading profits by faster, mechanised trading algorithms. At the same time, I understand my limitations in taking on the algo traders head on, on their turfs. As various warriors over the centuries have advocated, prudent wisdom demands that you do not participate in every battle that you are provoked into. That is the surest way to exhaust your resources and as ancient wisdom enlightens us - fatigue makes a coward of all of us. A wise General will fight a war which he thinks he can win, and will hit his enemy where it will hurt most. The same is true of algo trading. For what else is trading, if not a financial war?
Understand your opponent
The first step to beating your opponent is to understand your opponents modus operandi (MO) and his standard operating procedure (SOP). An algorithm is programmed with a primary objective of getting it's orders filled ahead of other traders. Speed is the primary MO here. Ever noticed how you were at the top of the bid / offer heap on the best five list on your particular counter, a better bid / ask trader comes along and betters your prices? Before your fingers can even leave the keyboard after modifying your bid / ask spread, your opponents rates have bettered yours in a nano second ! Nobody can be THAT fast, that's your first thought. You have probably just encountered an algorithm trade mechanism at work. Now you have one of the three options available to you - keep bettering your bid / offer prices so you finally get filled, risking getting filled at adverse prices that leave nothing on the table. Or, stop competing in this mindless game and stick to your guns and risk a trap wherein one leg of the trade gets filled and the algo wont let the other leg get filled - that's the ultimate horror of a spread trader or a scalper - a naked open position that refuses to get squared, involving capital, stress, time and opportunity loss, if not an actual loss too. The third option - quit and go home. Surrender to the system and withdraw from the fray, let go of trading opportunities for the want of competitive advantage.
Know yourself vis-a-vis your opponent
While many would argue that speed of execution is THE determining factor for a micro trend trader, and since algo traders rule that domain, micro trend trading's days are numbered. I would beg to differ. While majority of algorithm systems aim to get fast fills and weave and out of trades in the shortest spans of time, few programs aim to take directional views. And I mean directional views in micro trend trading. How does an average algo system determine where the price of counter ABC encounter overhead supply or drawdown support? What algo systems are primarily doing is fulfilling the objective aspect of trading - order fills as their MO. What about the subjective part that is a matter of human discretion and intellect honed by years of experience? My experience says algo fail to perform as well as humans when taking strategic decisions whereas they outperform their human counterparts on the tactical front. So, while you shouldnt be trying to beat an algorithm system on it's modus operandi (MO) front, the same is not the case with the standard operating procedure (SOP) front. Algos like almost everything in life do have a chink in their armors.
All algo programmers are reconciled to the fact that orders will be fed into the system (mostly DMA - direct market access via servers) like a chain. Therefore, the quantities in the order chain are never constant. There is a "scaling" involved. For example, if ABC is bought at 100, the system will buy 50,000 units, at 100.25, it will buy 25,000 units and at 100.50, it will buy only 10,000 units. The algo system also needs to be able to sell these units bought to be able to book profits, or the trade becomes an open risk, naked position. So at some point, the system will stop buying and start unloading. The system is also designed to unload as rapidly as it buys so the last lot of units bought are even acceptably sold at a minor loss, as long as the bigger chunk is booked at profits. This "tail end" of the algo system is what a manual trader needs to exploit. If you stare intently at your screen, you will see a pattern emerging here - just watch the price, volumes, open interest on every order filled, super impose these against the ATP and previous close. That tells you where your algo competitors are setting their limits beyond which they will not go. Make that your domain of choice, your kingdom where you rule the roost. Arguably, the quantum of trades filled at those price points will be smaller than the window price where algos are functioning but hey, war is never a perfect proposition !
Have a profitable day
Vijay L Bhambwani