Who decides which way price goes? The market, i.e. the buyers and the sellers. Say price is currently at X, and <...> there are orders at X+1 and X-1. Equal numbers of orders, or unbalanced, it doesn't matter, you pick. What determines whether price goes to X+1 or X-1 next? If the buyers exhaust all the supply of the sell orders at price X, the price goes to X+1. If the sellers exhaust all the demand of the buy orders at price X, the price goes to X-1. Also, why does it skip over numbers sometimes? Because the incoming orders from the buyers/sellers sometimes exhaust the supply/demand on several price levels instantly. Let's say the current price is 100, and the supply is: 100 shares at $100 200 shares at $101 300 shares at $102 400 shares at $103 Now, if I am a buyer, and I place a market order to buy 601 shares, and that order is executed, the current market price (as determined by the last trade price) instantly jumps to $103. Conclusion: DOM movement is random and provides no edge I don't see how you arrived to the conclusion by just asking questions. This is like asking someone about how the internal combustion engine works, and without even waiting for an answer, concluding that it can't possibly be used to move a car.
I think you're being a bit rough on the OP. His question indicate that he understands some of the complexity of keying off the order book. At least, his questions seem more sophisticated than your rather simplistic responses. When we look at the order book (I'm thinking of a central book such as Globex rather than a composite such as Nasdaq Level II), it looks like a lot of stuff is happening all at once. But at computer speed, updates arrive one at a time. As each update arrives, we have three basic questions to consider: a) buy b) sell c) wait Obviously, we have to decide size as well if our choice is a) or b), but most experienced traders can handle that. But reaching that buy, sell, or wait decision in the tiny fraction of a second we have isn't so easy. It's certainly not as easy as counting up the size at the inside 'n' levels of bids and offers. That's why I subscribed this thread ... to hear what other have to say (if anything).
That's a good analogy ... It's a really big poker game, but I believe there are "tells". They're not necessarily easy to spot, and they can change. My computer has the ultimate poker face. I never know what it's thinking.
I mainly scalp the ES, and sometimes looking at the Dome can give me an extra tic. I do have patterns that I watch for on the Dome, but often times they happen within one second and disappear, I will get a bias that price will go in my direction for 15 seconds or so, but for the average retail trader, trading the futures, I would suspect one tic shouldn't be an issue watching this for any length of time. But it can be helpful if you are trying to get out at a target and price is bouncing on your price and can't seem to get filled, I often watch to see if volume is going to change the other way so I don't lose more than a tic since I can't get filled on my target. I study price levels of volume and not total volume for five levels. When I see all five price levels on bid/ask go under 1000 orders, "big boys" are not playing the game, you can see this especially just before a report comes out or high volitilty, but if during the day and three of the closest levels on say ask side are under a 1000, and bid is just obe level under a 1000, my bias is price is headed lower. If I see during the day volume strink up, I reduce my volume asap. I sometimes see a "rotation" where I can judge price will go most likely go one direction for 4 tics. I see ask/bid bouncing at same price for 30 seconds, ask side might be 300 and bid side 900, then ask side increases whereas bid side stays same at 900, then bid side starts reducing faster, most like I will sell quickly before price goes down a level. I have watched a couple acquitances just trade off the Dome and no charts, but they have a knack and can't seem to be able to explain cause volume is constantly changing. I need charts, and the strain is huge by just watching numbers. But whatever floats your boat.
I have been thinking about this for a while and meaning to crunch some data but have not found the time. It seems like there should be something interesting in looking at book imbalances. I would think if market order flow is random (or uncorrelated with the imbalance) then price should move to the thin side of the book. And if the imbalance does predict future order flow then that is good too. Of course that doesn't mean there is anything left in time frames I can get to or after costs but still seems worth a look. I could see hidden liquidity being a problem, but you can just run tests with the visible portion and find out. I have found a couple studies on this, here they are. Still have not gotten to reading past the intros but they both seem to say there is at least some very short term predictability. http://snde.rutgers.edu/pubs/[31]-2008_QF.pdf