I assume that trades which happen in between the prevailing bid, ask are not recorded to the delta and are viewed as not having any MD significance?
I'd like to say thanks to all the excellent posts, sharing some great ideas and info about MD/time sales....comparing the >99 to the <5 lots on the es was a great idea and very helpful. Going back over about 10 days I found several times the <5 crowd diverged from price at the right time but once price went their way they seemed to fade it and go back to being wrong. For the most part the >99 was a better bunch to follow as most of you implied it would be. I've been plotting the accum. delta of >99, drawing channels and watching for breakouts. http://www.charthub.com/images/2006/10/30/2_mintest.png This chart shows a false break this afternoon where the <5 crowd was doing the right thing. I lost a couple ticks before I figured out it wasn't working so reversed to the previous vol based POC. Hoping someone can comment on the situation. May just be one of those things, nothing will always be right. Playing the breakouts has worked nicely lately though. Wish I had something to offer this crowd, sure am getting a better grasp of things through all the informative posts. The morning trade mentioned in a previous post continues to work. Using a 1 min chart and using the open vs. vol POC of previous day relationship. I enter on first delta surge away from the POC. Usually a reading of +-1500 to 2000 on a 1 min bar. Price poked into the vol POC, almost filling the gap before the delta got me in short. Haven't done this long, usually enter this trade in the first few minutes.
Just the kind of interesting picture post i like to see in the MD group thanks... cj... HAVE STOP <img src=http://www.enflow.com/p.gif> WILL TRADE
ticktrade, I'll echo cj's comments - excellent post. Your chart clearly shows what I made reference to in an earlier post, and that is, the "retail" guys are not necessarily the cliff jumping lemmings that conventional wisdom would have you believe. Whatever, I guess the important point here is that there is tremendous value in watching what the big guys and the little guys are doing. When the two are in-sync everybody is agreed as to near term direction and the market should faciltate trade very well. When they're not, it points to a conflict of opinion which may serve as a warning, or indeed create an opportunity. Ticktrade, perhaps you can help me out here - in MD/IRT how do you apply an indicator to indicator? For example, your MA in the VB. I haven't figured this out yet. Either I'm an idiot or it requires the Pro version (or both). Looking forward to more great posts from you. bolter
paul, I believe you are correct, although I'll confirm this with Chad. It's not that they don't have any significance but it is impossible to infer whether it was likely a buy or sell transaction. On a liquid market with a decent tick size (eg: ES) the spread will be a fairly consistent 1 tick, so this will represent a small proportion of the total volume. On thinner markets, or markets with a smaller tick, obviously it could be a much greater proportion of the total volume. I suspect it could have a material impact on MD output, but I have no evidence. From my experience however, I find MD is less useful on these types of markets. bolter
A little while ago, I had a look how many trades were falling between bid and ask using IB data feed. It was pretty much as you say - the number is relatively small in highly liquid, high volume markets with a 1 tick spread and significantly larger on thinner markets. If I remember correctly the figures were something like 1% on ES, K200, a couple of percent on ESTX50, 7 or 8 percent on DAX and ER2. Maybe something like 12% on HSI. The winner was crude oil on IPE (or ICE) which frequently gave figures of 20%+ which would be big enough IMHO to seriously affect delta. Again these sorts of figures could well be significantly influenced by data feed vendor and exchange software. I made the decision to just discard these trades when calculating delta. For large spreads, one could possibly allocate them to trades at bid or ask according to which side they were closest to, but it was not clear to me that this actually represents anything meaningful.
dcraig, Great research. That's the sort of stuff you instinctively know to be the case but never actually get around to proving it or indeed capturing the stats. Well done - thanks for sharing that with us. All the best, bolter
These figures come from bid and ask as reported in Level 1 feed. They may or may not be the same if best bid and ask from Level 2 stream is used. I havn't tried the latter. One market that is a bit odd is SGXNK. As we know, tick size is 5 pts, but trades are frequently reported at other than a multiple of tick size. These will fall between bid and ask. IB say the trade prices are correct. I don't fully understand this, but I assume these trades are not going through the book ?
I see the "part way" prices on the live feed too dcraig. There is something wrong with IB's datafeed. As with bars that don't include the true high or low, when you do a refresh the data is corrected.