Intraday Market Statistics Wish List?

[Mike Condra]

Suppose you could examine all market activity as it occurred.

What intraday statistics might you collect about individual stocks, exchanges, sectors and the market as a whole?

[Enlighten]

Jee I suppose some might consider that a loaded question
I am sure I could come up with a thing or two but would need to check out esig and their addons to make sure its not already here somewhere.

Would this wish list be for esig next full version, or for another addon product?

[Mike Condra]

Michele,

The thought is that arriving tick data could be examined in more detail, to create new fields of interest to eSignal subscribers.

VWAP (volume-weighted average price) was suggested. An approximation to "official" VWAP can be had by summing size*price over all trades and dividing by current daily volume.

It's possible to collect statistics about the size and time distribution of trades, either for single stocks or collections. If those numbers could be graphed or compared with earlier ones, they would start to have meaning for spotting trends.

As ideas in this area surface, please share them.

-M

[Enlighten]

ah ha I see.... thanks and I'm glad the bb is working again

[bigtee]

how about
advancing issues, declining issues, adv- declining issues, adv vol, decl vol and adv vol -decl vol for the NDX? For the SPX?
For specific ETF's using the stocks that make them up like SMH?

How about a TICK index for the NDX and SPX?

Of course there is also the vast array of market indicators at QCharts/LiveCharts, like most up in price, most down in price, Up on above ave volume etc.


A putcall ratio would be nice.

[its1111]

Wish list

I would enjoy hearing more peoples thoughts on this, but I can tell you that it is very imporatnt to also watch the $tick which is basically how many issues are advancing or declining there are patterns. I also watch the premium levels and the $ticki.
The prior thread watches the Vwap and you can get that on the old charts. It is extremely important to know what price the funds are buying at. If other people have more I would love to hera.
Also it is smarter to use the $adv minus the $decl. because they update every 6 sec. instead of 30 seconds like the $add.
Hope this helps.
Mark

[theplumber]

Re: Wish list

Originally posted by its1111
I would enjoy hearing more peoples thoughts on this, but I can tell you that it is very imporatnt to also watch the $tick which is basically how many issues are advancing or declining there are patterns. I also watch the premium levels and the $ticki.
The prior thread watches the Vwap and you can get that on the old charts. It is extremely important to know what price the funds are buying at. If other people have more I would love to hera.
Also it is smarter to use the $adv minus the $decl. because they update every 6 sec. instead of 30 seconds like the $add.
Hope this helps.
Mark

Something I've done (using an Esignal competitor) is to collect all 500 S&P symbols and make my own TICK reading as well as AD reading. Of course this depends on how fat your pipe is as well as horses under your hood. Of course stats and bots are pandoras box on Wall Street as most would be surprised as to what some people have "sniffing" around looking for "anomolies". One bot I use was of use today, a big seller of futes (Goldman) out of the gate wasn't hedged like a normal sell/buy by the dealer. This speaks louder than you can imagine if you sling size, the real seller sold 1300 CARS total and the bids disapeared, so I'm not the only one who has thought up this kind of statistic to watch for. My same program also looks for a bond to stock move (and vis versa) and today it hit, the seller was going into bonds. This little peice of info ultimately leads to easy shorts, or just short any and all rallys. Basically any tape reading skill can be automated to a small statistical program to look for anomolies in just the price action.

[its1111]

intraday stats

Mr. plumber
I was curious to know how you knew that Goldman was selling and was not hedged. Did you listen to a sqawk box. Do you know they weren't buying options. I only ask this to learn more, as I am on a never ending search about the markets.
I have read many of your posts and it is obvious that you are very smart when it comes to trading. So any insights into your screen set ups or indicators would be appriciated.
Mark

[theplumber]

Re: intraday stats

Originally posted by its1111
Mr. plumber
I was curious to know how you knew that Goldman was selling and was not hedged. Did you listen to a sqawk box. Do you know they weren't buying options. I only ask this to learn more, as I am on a never ending search about the markets.
I have read many of your posts and it is obvious that you are very smart when it comes to trading. So any insights into your screen set ups or indicators would be appriciated.
Mark

Yes I use sqawk but don't know who's specific service b/c I'm part of a larger group. The hedging conclusion by me is my own propietary work I've done and is based on how the street really works (no secrets, just stuff you won't find in any books, ever) . One of the best programs I have is spotting program trades. This is the reason I collect all 500 symbols so I can have realtime info, the saying is markets don't move, they are moved. J. Succo wrote this peice back Dec 12 to explain better than me how to spot a program trade and how the street really works, enjoy.

Yesterday’s strong market perplexed many. The morning’s economic data if anything was weak: bonds by the end of the day were stronger. After the data was released S&P futures were trading lower and it looked like stocks would be under pressure. But unexpectedly stocks were firm at the opening and gained strength as the session wore on.

It seemed that buyers completely ignored the economic data. But perhaps they had already made the decision (and the commitment) to buy even before the data was released. If this was the case, the primary buyers would probably have to be a small select group, maybe even one participant. The previous day’s close may give some clues as to why this may be so.

At 2:30 the previous day the market was at its weakest. All of a sudden it stopped going down and began rising precipitously. Analyzing the tick data at this time reveals a very significant 1500 point swing (from a -1000 to +500) at precisely this time. The selling was almost instantaneously overwhelmed with buying. It seems someone knew something.

I have written in another piece of how in the 80’s and 90’s there was a great effort by mutual funds and their consultants to reduce commissions they paid wall street brokers. Wall Street, being a little smarter than their customers, devised strategies that took from Peter to pay Paul: they acquiesced to lower commissions at the expense of market impact.

One of the strategies that developed as result of this process that was designed in an attempt to mitigate this new problem of market impact is called a “blind program”.

A customer who wants to buy a list of stocks (and/or sell a list) would submit to a select group of brokers a statistical description of the program to “bid” on. The broker does not get the actual names of the stocks. The bid price is in a commission amount.

For example, let’s say on Wednesday three different brokers got a call from BS Mutual fund for a request to bid on a blind program for the next day. BS sends each broker a sheet listing things like dollar amount to buy and dollar amount to sell broken out by sector; average, mean, and maximum and minimum market capitalization of each sector; liquidity statistics of each sector; etc.

Let’s put it this way: enough statistics are revealed to the broker that even though they do not have the names of the stocks, they can very precisely determine the amount of risk involved in executing the program. Each broker then bids in cents per share what they are willing to charge to do this program. The risk is that they promise the customer the closing prices of the previous day.

Why would a broker promise the previous day’s closing prices: this seems on the surface a great deal of risk to take? I will tell you that statistically there is risk, but maybe not as much as you think. Some stocks they will make money on and some they will lose money on, and they are working against themselves as they buy stocks that they owe offer prices on and sell stocks that they owe bid prices on. But they also get a great deal of extraneous information on which they can make a great deal of money, the primary one being dollars to buy and dollars to sell.

So on Wednesday at 2:30 a few brokers received a statistical sheet that among many other things told them that the next day the customer was net buying a great dollar amount of stocks. That information “somehow” got into the market right away.

An analysis of the tick data from yesterday reveals a pattern consistent with program buying. Program buying normally represents buying from a small number of participants, not buying from a large diverse group of investors.

I do not know for sure that this was the case, but the pattern fits the data. Regardless, it was an opportunity to explain one of the obtuse tools used by wall street.

[Enlighten]

I am supposedly subscribed to this thread and have received the emails on the replys up until the last two messages posted , there are no emails that I have received here for the prior two posts ....is there a problem? I did not unsubscribe here....

[ScottJ]

Michele,

We were having some server issues earlier so that's likely why you didn't get email notifications. Should be OK now. If not, please post again.

Thanks.

[theplumber]

Of course the best statistics to make money on are your own using somthing as simple as excel, you just have to be creative in doing your queries. Because all my services are tax writeoffs I can afford services like this one http://tradeplotter.com/index.asp
but you can do something like what they do on your own. For example this is what the founder said this morning and you can also get him here every morning from my friends at http://www.minyanville.com/gazette/

[BKuerbs]

To turn these figures into meaningful Statistics you have a long way to go:

• What do you want to do with one-time events?
• Averages in the financial markets can be pretty misleading. I'm missing the variances.
• But above all, I'm missing the test for statistical significance: You have to test, by an appropriate statistical test, whether the returns after any event you defined are significantly different from the occasions where the event did not take place.
• Statistical significance does not mean economical significance: read it will not always translate into Profits. The reason are the usual suspects: bid/ask spread, commission, and slippage.

Without these additional numbers, your figures are meaningless.

Did you read on the subject "Data mining" and "Data snooping"? There is a lot of literature to be found in the internet covering these subjects.

Regards

Bernd Kuerbs