Hi bolter, I think my mistake is to compare the delta with and without the odd lots filter. The small guys' effect on the overall delta is so small that I don't see any difference. However if I only compare the small transactions with the large transactions, I see the difference you and MK describing. It's sad to see the small guys (like me) get squeezed by the big guys some times. As to my studies on delta persistence, there is not enough data yet and I don't want to give people the wrong idea and hindsight. But one thing I would say, if I am looking for evidence of persistence, I should probably use volume chart instead of tick chart or time chart because delta is based on volume... this way we are quantifying apples with apples. Thanks for everybody's generous input. Regards, William
Hey guys. I have an extremely basic MD question I was hoping someone could clarify for me. I was looking at the footprint graphs and realized that they do not make much sense to me. If market delta is the net buying or selling for any given price,time, how is it possible that market delta is anything but 0 for any given price and any given time? The graphs show red and green squares depend if there are more buys or sells at a price/time, but for there to be 10 buys say at price 1050.50 at 11:30:01. does'nt that mean there needs to also be 10 sells at price 1050.50 at 11:30:01 to match up with the buys? how is it ever unbalanced such that it is possible to see more buying or selling going on? Thanks for putting up with my newbie question.
dood. buying is defined as a price going through at the ask. selling is a price going through at the bid. this is a standard convention in terminology. technically speaking, EVERY bought contract has a corresponding sold contract. that's a mathematical certainty. it can't be any other way but when trades go through on the ask, that is conventionally believed to be indication of buying pressure, since somebody is "paying up". the corresponding opposite is true on selling and the bid. anytime somebody wants to get IN right away, they have to pay the ask. that is the only way you can be guaranteed instant execution. so, taking the ask is seen as a "buy" even though if the guy bought on the ask, somebody else SOLD on the ask. but the seller who was filled already had a limit order to sell on the ask. so, he was not entering on an immediate trade, he was filled after some period of time after offering previously (since he is next on the offer queue). the only thing that moves prices is up is when enough people "pay up" on the ask, so as to take out offers on the ask, which moves the ask up 1 tick... and so on... stop runs of course, and automatic closes of positions due to margin calls (which is when you see the most dramatic moves in a market) accelerate this process. and of course, people lifting offers cause they see the price rising fast, and they lift an offer to place it higher. btw, bolter - your MP thread was one of the best EVER. loved it. i don't use MD, and probably won't, at least not for some time. but I am greatly enjoying this thread.
Nice post above by Whitster, and a good thread overall. I'm re-posting something that I said in another thread but applies here as well: -------------------------------------------------------------------------------- One caveat though is the MD premise itself, which says that it is the large aggressive traders that move price and that buyers and sellers are going to market, not sitting in the que with limit orders. The MD assumption is that ask traded volume is always buying and bid traded volume is always selling. However, there are big players who don't go to market, but sit in the que; and some of the biggest players use a combo of hitting and sitting. Hitters vs. sitters - understanding their different motivations and subsequent behavior is the key. ------------------------------------------------------------------------------ I would add that although I do use filtering for size, it is far from a panecea. It's fairly well known that someone moving real size can easlily chop up their order into many smaller orders and have them go through almost simulataneosuly giving the impression that it is a bunch of small traders.
paul, One point of clarification from whisters post for those not familiar with tick data. Data vendors supply tick data in a format such as: date, time, price traded, volume, bid, ask with 1 record for each "trade". So by examining the price traded relative to the prevailing bid and ask one can infer that the trade took place at the bid, the ask, or somewhere in between. bolter
I gather volume stats on ES Futures and SPX cash every ten minutes from the open for 20 days... (its always a running forward sum) So once i have done that i then take the 20 days worth of data and run them thru a Query to find out for <b>EACH</b> 10 minute period in the day what is Normal Volume, what is 5% less than normal volume, 10% less than normal volume, 20% less than normal volume, 25% less, etc... and on the other side what is 5% greater than normal volume, 10% greater, 20% greater, 25% greater, etc... then print it out as a report that has every 10 minute period of the day showing what is normal, +/- 5%, 10%, etc... agian so once you know this you have an idea of the potential for the day and where you stand right from the open and for every 10 minute period there after... similar to comparing the futures volume now to the the 30 day's average volume but that is not accurate since the day's volume is so variable in growth throughout the day... and you cannot do this for the SPX volume versus a posted average volume since there are no supplied volume figures for that... anyway once you know if both or either of the volumes (SPX & ES) is greater or lesser you know if Outside paper... big players, institutions, etc... are participating or not and you can make a decision as to it being a trend day or normal day or chop or refinements of that... with less effort and a better guesstimate hope i made this clear... its not rocket science... just takes collecting the data with an Excel DDE & a Dbase setup or other coding method for the SPX volume and then running some Standard Dev queries... The ES futures volume is available to parse all the time, at EOD, or collect as it is executed too... cj... HAVE STOP <img src=http://www.enflow.com/p.gif> WILL TRADE
Top of the morning to you Edge. That is exactly what I was driving at in an earlier posting ... well done. Trade Maven Pro offers volume collection on a similar bases to your very clever excel programme. I ran it for a while and it really was the dawning of my understanding of vol / price. Also watching the big boys at the pre open will give good insights into the coming day. Something else for Thread Readers to consider is to watch trades on the bid print through on a rising tick or trades on the ask print through on a falling tick. Pay attention to the strike size ... ie ignore the minows.
One more thought for all. Vol & price are two different animals. Related and co dependant for sure, but never the less, different. In order to compare them, try applying a stochastic to both in the same window. Same period, you my want to smooth them a little. I think that you will be fascinated. Sometimes you will see vol run far ahead of price and vice versa. Vol can be running at 80+ and vol stubbornly remains at 20 or less ... and guess what happens next. If you can think of something better than a stoch, I would appreciate your input. I have fiddled around, but stoch seems to fit just fine. Anyway, another day another dollar. One thing is for certain. If you keep feeding this thread with positive input, we are all going to be the wiser by Friday. Good luck to you all.