The following idea(s) came to me after reading a bunch of threads. Expert feedback would be appreciated: STARTING ASSUMPTIONS - The majority of bids and offers are placed by a small number of market making (MM) algorithms. As a starting point, I usefully imagine that a SINGLE MM is responsible for all bids and offers. This assumption is almost certainly not accurate but for the sake of constructing a trading strategy, I choose to hang my hat here for the moment. - Therefore, when a trade is executed at the offer, I assume the MM is short; and when a trade is executed at the bid, the MM is long. - Using Bloomberg Terminal data, I have calculated what I regard as being the MM's running P/L position, in addition to the value of their accumulate net long or short position. Usually these net out to a modest profit (thanks to the spread). - The MM inevitably ends up with an accumulated position which is "offside" (e.g. because if the market is going DOWN, bids are being heavily hit and the MM is forced to accumulate a net long position in a falling market). MM BEHAVIOUR? - From what I can gather, MMs face limitations in terms of the size of their accumulated net positions (i.e. "maximum inventories"). Therefore, running the calculations I mention above may allow me to identify when the (assumed) single MM is heavily overloaded with inventory (either long or short). I further assume the MM will do one of the following things in such a circumstance: - if overloaded with longs, they will stop placing bids (or if overloaded with shorts, they will stop placing offers). This approach may be counter-productive as it may simply send the market spinning further "offside". - if overloaded with longs in a falling market, the MM might place MORE bids in an attempt to encourage other participants to buy, thereby giving the MM a chance later on to offload at a more favourable price. (NB: Perhaps some or most TA patterns can be attributed to MMs as they attempt to entice other participants, simply to generate more liquidity to offload their inventory?) - if overloaded with longs in a falling market, in the most dire circumstances the MM could "cross the spread" and sell some of it at market. The trouble is this will make the market drop like a stone and the MM could get rekd. POSSIBLE STRATEGIES? In view of the above, possible strategies may include: 1. When I suspect the MM is overloaded (to be determined by examining many weeks of data), place my own trades in the expectation that the MM will take action to flatten their position (e.g. "puke" out their position). 2. When I suspect the MM is overloaded, wait until the "closing bell" (if there is such a phase for the chosen market) and anticipate the MM will be forced to liquidate at that time so they aren't holding a net position overnight. WHICH MARKET? Feedback on any or all of the above would be great. If there is any merit to my thinking, I would also appreciate suggested markets to focus my efforts on. Kind Regards,
"- Therefore, when a trade is executed at the offer, I assume the MM is short; and when a trade is executed at the bid, the MM is long." Perhaps you are wrong and MM is exiting. Unless you can read order flow and the Dome well enough, or Quantum Physics for projecting price based on where it just came, will be most difficult. Instead of trying to find flaws in HUGE VOLUME traders, I rather try to discover how they entered. Fighting against big volume for ticks? What happens if more big volume joins in and both move the market, then you get snowballed. For myself, finding flaws in retail is easier, get in couple ticks earlier and reverse when they are entering and wait for big volume to push other way like in ES to cause tight stops to be hit and reverse again, like in trend line breakouts. There is nothing magical about some trendline containing or breaking out price unless big volume allows it. It is like Fib retracements, they work cause enough size allows it to work.