Register for free to join our community of investors and share your ideas. You will also get access to streaming quotes, interactive charts, trades, portfolio, live options flow and more tools.
Site Update
The class on "Scientific Measurements" is now available in the Geek Academy. If you think "precision" and "accuracy" mean the same thing, you may want to go through that class...
Hypothetical Account Update
Our order to buy XLE at 56.09 did not fill, so there is no active trade this week. XLE gapped up and opened at 56.67 on Monday. It reached a low of 56.19 which was still 0.10 above our limit price, so we barely missed it. Better luck next week!
Site Update
At long last, the Bar Pattern Analysis class is finished! Here's a link to the news item on the site:
http://www.market-geeks.com/news.php
Hypothetical Account Update
The new trade for next week is posted on the home page. This one is a buy of XLE, the energy sector ETF. The hypothetical account will attempt to buy 173 shares at a limit of 56.09 on Monday.
Hypothetical Account Update
Gee, a trade that actually made money? LOL! XLU hit neither our stop or our limit this week, so we assume an exit at the close of 30.24. We were short from 30.46 so this gives us a profit on our 317 share position of (.22)(317) = $69.74
Zippy!
I am updating the site this weekend and will post a new trade sometime tonight or tomorrow night.
Hypothetical Account Update
XLU fell pretty much all day today, so now our trade is ahead by about $48.
Hypothetical Account Update
Based on the intraday price action of XLU, our limit order to sell at 30.46 would have filled in the first half hour of trading as the price opened below our limit and proceeded to climb through it. The account is short 317 shares of XLU, and as of today's close the trade is down about $63.
Now that we know what our entry was, we can assign our stop and target prices. They are to be 1.05 points on either side of the entry, so we have:
Stop: 31.51
Target: 29.41
Neither was hit subsequent to our entry today.
Site Update
I've added a new news item to the site about the upgraded BPA tool. http://market-geeks.com/news.php
As I mentioned there, when I finished with the upgrade I had to re-test all the patternsets with all the securities again, so there are some changes to the info on the site. The best patternset for a given security has changed, and all of the top-ten individual pattern lists have changed. These can be found in the Geek Academy.
When I ran these tests (for the FOURTH time since I created BPA, grrrrrrrrr....) I made some interesting observations. The patternset scores are generally higher, and so are the top-ten pattern scores. Several of them show return lifts of over 2% per trade!
Also, the best patternsets for each security are somewhat less randomly distributed now; they are more clustered, which indicates a higher degree of consistency in which patternsets perform the best over a range of securities. This may indicate that there is some "meaning" or predictive power inherent in these patternsets.
Here's the distribution, showing the securities for which each patternset scored the best:
#1: XLB
#2: XLV
#3: XLP,XLF
#4: DIA,XLI,XLK
#5:
#6:
#7: SPY,QQQQ
#8:
#9:
#10:
#11: XLE
#12:
#13:
#14: XLY, XLU
You can see the clustering in the first four sets, particularly sets #3 and #4. These are the patternsets that just use the changes in the closing price as parameters. Interesting.
I also noticed that patterns from set #4 and #14 were very common in the top-ten lists, but I haven't done an actual count. One of the future research projects for the LAB will be to put all the data from the top-ten lists into a database, right down to the specific individual pattern characteristics, and see if there are any specific signals (like an increase in the low accompanied by a decrease in volume or whatever) that show up frequently in top-ten patterns. In other words, that project will try to answer the question, "What do really great patterns have in common?"
But that's in the future, and I can't do ANYTHING until I've got the Geek Academy finished. The poor students are getting rained on with no roof on the building and only half the walls up, LOL! The BPA class needs some major revision now because of the upgrade, but that should only take a few days. That's the next step....
Out here!
Hypothetical Account Update
Wait! Don't ring that opening bell just yet! Wait for meeeeee.....
Whew, I managed to finish the upgrades to BPA and still squeezed in this week's update in time for the open. Yay!!
This week's trade is posted on the home page: http://www.market-geeks.com
As you can see, it's another short trade in the utilities, but this time we have a PTRL score of over 2%. Also, this will be the first trade generated by "BPA 2.0", so we'll see if those constant stopped out trades continue or not. You'll notice a change in the target and entry prices; they're not prices. They are listed as point values from the ACTUAL entry price (which we don't know yet).
Stay tuned...
Oh I see...the trades all seem backwards, LOL!
I think a lot of it was due to some whacked out logic that I originally had for the optimization routine. I fixed it once, but am working on a second fix now that I described this past week.
Also I'm using those original 13 patternsets, which are all probably awful. They're very similar to one another, and you can probably tell that it only took me about an hour to come up with them.
They're good enough for testing the BPA procedure though, and once I get the bugs out of it, finish writing the class (which is helping me to identify the bugs) and finish the rest of the Geek Academy, I can then start using BPA for its intended purpose; testing a bunch of possible patternsets to find ones that have predictive power.
After I've tested a few hundred patternsets with BPA and found some really good ones, I'll probably look back at the scores of these original 13 sets and laugh at how terrible they were. At least that's what I'm hoping, LOL!
By contrarian I was kinda joking but I meant that it seems like you woulda made money on every trade now if you had just been on the other side of it. Long instead of short, etc.
Hypothetical Account Update
We were stopped out of the XLU trade today at 31.38. This is a loss of about $104. I will update the trading history during the weekend.
I'm not sure if I will be able to post a new trade this weekend because of the revisions I'm making to the BPA optimization routine, but I'm shooting to be done by then.
Site Update
I have good news and not-so-terrific news.
The good news is that I will be able to add FOREX analysis to the member's area sometime soon.
The not so great news is that while writing another lesson for the BPA class, I have identified YET ANOTHER flaw in the optimization routine. This isn't really bad news because I'm kind of TRYING to find the problems with it, but in another way it bites because I have to fix my spreadsheet and then retest every single patternset with every single security. With 14 patternsets and 12 securities, that's 168 tests, a process that will take at least two or three days. And then, I will need to make revisions to some of the lessons in the BPA class to reflect the revision.
This time, the problem is that I was pegging the planned entry, stop and target prices to the previous close. This seems to make sense, except when you get a really good entry on an actual price bar, and then the stop price in the trading plan makes no sense.
For exampple, say the last close was 50.00, the local range is 4, and we're using this trading plan: {+5%, +50%, -25%}. This would translate to:
Entry: Buy at 50.20 (or below)
Target: 52.00
Stop: 49.00
Well what happens if the next bar opens up at 48.90? Our entry order would fill, and we'd be in at 48.90. Great price! Now what's our stop price? Oh...49.00? Wonderful. Our order is to sell immediately if the price drops below 49.00, and the current price is 48.90. Which means we're stopped out before the trade even starts? Huh?
This makes no sense, so in order to fix this, I really need to peg the stop price to the ACTUAL entry price, not the last close. For uniformity, I'll also peg the target price to the actual entry as well, although this isn't actually necessary. So...the planned entry will be pegged to the last close, then we wait to see if there's an actual entry, and if so we determine the stop and target prices based on that actual entry price. That will ensure that the stop and target always make sense.
Right. THIS should be fun to fix.
Glork.
Hypothetical Account Update
Based on today's intraday activity in XLU, our limit order to sell 315 shares at 31.05 would have filled between 10am and 11am. Our stop of 31.38 was not hit subsequent to that, so we are still in the trade. As of today's close, the trade was ahead by about $57.
Hypothetical Account Update
All of the analysis scores came out really low this week, and we ended up with a short trade on the UTILITIES of all things. The new trade is posted on the home page. Note the low PTRL score of 0.26%. Yeesh.
There was only one individual pattern alert in the members area as well this week; a short trade on XLF. There are usually several alerts based on individutal patterns.
Interesting phenomenon at any rate. It's as if every security is signalling a "ho-hum" week ahead. Even if the prediction isn't accurate, the interesting thing is that all the different patternsets used to analyze the different securities are giving similar results simultaneously. Not a very rigorous observation of course!
Hypothetical Account Update
Our XLY trade didn't hit the stop or the target price this week, so we assume an exit at the close of 33.66. This results in a loss of about $52.
Hypothetical Account Update
Well, our XLY trade narrowly avoided being stopped out again this morning, reaching a low which was 8 cents above our stop point. It moved up later in the day though, and our trade is now up by about $34.
This hypothetical account is kind of like a "meta-experiment" with BPA to see how it works in the market. In the true spirit of science though, I'm not trying to prove that it works. I'm trying to poke holes in it and see if I can show that it DOESN'T work.
Science is an ongoing process of eliminating hypotheses, until you're left with the ones you can't seem to eliminate. This set of resilient hypotheses then form the basis for our current theories and paradigms of how the world works.
You might say that science is a series of failed falsifications...
ANYway, until next time!
Cap
Almost seems it might be a good contrarian indicator, lol.
Site Update
Lesson #5 is now (finally!) up in the Bar Pattern Analysis Class. http://market-geeks.com/academy/research/bpa/bpa5.php
Three more lessons to post. Will this BPA class EVER be done???? It seems like I've been working on it for a month.
Hypothetical Account Update
Almost stopped out of the XLY trade today, but not quite. Our stop is at 33.52 and the low for the day was 33.58.
Yo capitalist: jump on this little piggy tomorrow. ITWJ gonna be monsta. Hope you are fine?
Hypothetical Account Update
Base on the trading activity in XLY during the first hour of trading on Monday, the account purchased 290 shares of XLY at our limit price of 33.84. The trade is down about $17 at this point.
The account history is one of the items in the "Directory's Office" and is located here: http://market-geeks.com/directorAperformance.php
Hypothetical Account Update
Here is the trading plan for the coming week:
Security: XLY
(Consumer Discretionary Sector ETF)
Last Trade: 33.79
Pattern #11-2/16
PTRL= 1.48%
OPTIMIZED TRADING PLAN
Entry order: BUY @ LIMIT of 33.84
Stop loss: SELL STOP @ 33.52
Exit target: SELL LIMIT @ 34.33
Full analysis is here: http://www.market-geeks.com/directorDfree.php
Site News
I have finished updating lessons #1-4 in the BPA class. Lesson #1 now has a brand new overview diagram of the entire testing process.
Hopefully someone will tell me if this diagram makes any sense and if it helps to clear up what BPA does.
I won't be here tomorrow night, so I'll update the site for next week's trading activity over the weekend. Looks like the error that I found only cost about a week of down time.
Out here...
Hooray! All finished fixing the optimization routine in the BPA Scoring Engine. I've also rerun all 14 patternset tests for each one of the 12 securities I track on the site. That was a total of 12 x 14 = 168 tests. Each one takes a minute or two to run.
The scoring engine is a spreadsheet in MS Excel, but you can think of it as a big machine with two input funnels. Into one funnel I pour all the data from 300 price bars of some security, like QQQQ. Into the other funnel I place the detailed description of a particular patternset, like Patternset #14.
Two things come out the other side. First, the patternset receives a score for how useful it was in lifting returns for that particular security. Secondly, after every patternset is run, I get a list of the top ten performing patterns for each security, regardless of what patternset they came from. This is like an all-star team in baseball, where the best players from all the teams get together. The patterns are like the MVP's and the patternsets are like their home teams. So for example, Patternset #6 may suck with QQQQ, but an individual pattern that come from that set, like #6-3/16, may be fantastic when used with QQQQ.
Anyway, now I can get back to work on the Geek Forum. I think before I start on lesson #5, I'll try to redo lesson #1 with a less detailed explanation and maybe a diagram to show an overview of what happens inside the BPA Scoring Engine, and how optimization of trading plans fits in, etc.
Out here....
Just meat - no potatoes.
LOL, good slogan!
Hello CApitalist: hope you are fine, buzy I see. Visit my new board: http://www.investorshub.com/boards/board.asp?board_id=5388
Site Update
Ok, the BPA class will now have a total of eight lessons, and the first four are completed. The information that was originally in Lesson #4 is actually still valid, with some minor changes that I've already made. This was all the stuff about the meaning of an optimum trading plan, the notation for writing the plans and the use of the Percentage of Local Range (%LR) scale. So don't brain-dump that.
Site Update
Glork.
I've recently uncovered a logical flaw in the optimization routine I use to determine the trading plans. Disregard the last two BPA lessons; the ones about determining the optimum trading plan for a single price bar. I'm taking those down and re-working the syllabus.
When I first created the optimization routine in BPA, I was using a dataset with around 1500 price bars, and it was so complex that I didn't notice the flaw. As I was writing Lesson #6 to explain the routine, I used an example with only five price bars. That's when I saw the error.
The solution is just to use a simpler, more straightforward optimization routine which I have already developed. It will take me a day or two to rework my "BPA Scoring Engine" in MS Excel to incorporate the simpler routine, and another couple of days to run all the old price data through the new system.
So I won't be doing an update this weekend...everything should be running again by the following weekend.
Glork.
Hypothetical Account Update
And the hits just keeeeep on comin'...
Pretty quick trade this week; we were stopped out of XLI at 33.19 today, for a loss on this trade of about $94. Ouch!
I have already updated the performance summary on the homepage and the account history in the Director's Files.
Tune in next week!
Hypothetical Account Update
Assuming that we filled our order at the worst price during the first full hour of trading today, the account sold short 303 shares of XLI at 32.88. Our entry order was at the sell-limit of 32.76.
As of the close today, this trade was ahead by about $15.
Our stop price is 33.19
Our target price is 32.13
The plan is shown on the homepage: http://www.market-geeks.com/
A full account history is shown here: http://www.market-geeks.com/directorAperformance.php
Site News
Lesson #5 is also done. That one was pretty easy.
Site News
Lesson #4.
Done.
Thrills.
http://www.market-geeks.com/academy/research/bpa/bpa4.php
Site News
I've updated the "class syllabus" for the BPA class. There will be a total of 11 lessons, making that by far the most extensive class in the Geek Academy (as I expected). Three of the lessons are done, and it will probably take me another couple of weeks to finish off the other eight.
Once that's done I can move on to working on some of the more basic classes again. I hope to be finished with the entire Academy before the end of May.
Then it will finally be time to get started on the research projects that are backing up in the Members Area. Some of these are related to Bar Pattern Analysis, while other projects will be heading off in entirely different directions instead.
Anyway, that's the plan for now.
Cap
Hypothetical Account Update
Ok, the site is updated for this coming week.
The free weekly trade might look familiar, it's another short sale of XLI. That's because XLI exhibited the same pattern that it did last week; #4-16/16. That particular pattern is an easy one to describe. It occurs when the stock closes up four times in a row. Historically when that happens in XLI, the prognosis for the following week is that the price bar will have a longer tail down than up.
So here we go again...
Hypothetical Account Update
Well XLI didn't hit the target or the stop price during the week, so we exit the trade at the close of 32.85 which is 20 cents above our entry. We're down $61.00 on this trade, for our second straight loss.
Of course after only three trades, I can't really expect the law of large numbers to be in effect yet...
I'll be updating the account history here: http://www.market-geeks.com/directorAperformance.php
And I'll be updating the site for next week's trades.
Site News
I've posted the third lesson in the BPA class. This lesson begins a step by step illustrated project for the creation of Patternset #14. Have fun!
Hypothetical Account Update
XLI dropped a bit closer to our target on Wednesday, but no exit yet. The trade is up about $30.
I won't be online tomorrow night. See everyone Friday night!
or a few hundred subscribers, lol!
LOL, gotcha. Sounds like you need a worker monkey.
Hypothetical Account Update
It was kind of a bouncy day for XLI, but the price didn't hit our stop or our target, so we're still short 305 shares in this week's trade.
Hi Matt,
Originally I had planned on using some weekly price bars in the free area, and providing the members with daily price bar information. However, it takes me at least one full night at work during the weekend to update all 12 of the securities I'm tracking for the following week's price bars. Even though I have a spreadsheet set up to run all the calculations quickly, I still have to:
update each dataset by one price bar
load the new datasets into my spreadsheet
run a new analysis, one security at a time (automated)
tweak the resulting images a bit (security names & dates)
upload the new images to the website
create the summary table of all 12 securities (members area)
run another spreadsheet to look for "top ten" patterns
create a "top ten" summary table (members area)
create and upload 12 individual "top ten" analyses (members)
and a few other housekeeping things.
So it was pretty clear that there was no way I could even post daily updates until the site brings in enough revenue for me to work on it full time. So that's why everything's weekly.
Daily bars may become possible at some point with a great deal more programming and automation.
Intraday is basically science fiction at this point.
Have you thought about using BPA on an intraday basis? It seems to me that if it were to be hooked up to an automatic trading account whereby the transactions were done with precision timing by a computer running BPA to get the cues it would be more effective.
There are also a lot more technical indicators for minor price swings in intraday data than in daily HLC. I'm thinking of GOOG for example.
Maybe instead of .13% a day you could capture .13% a couple of times a day and yield higher returns. Of course you'd have to be trading with a large amount of $$ otherwise transaction costs will kill ya.
Or am I totally wrong?
Hypothetical Account Update
The hypothetical account entered this week's trade by going short 305 shares of XLI today at 32.65 as specified in the free trading plan I posted on Saturday. In real life we would have gotten a better price of around 32.70, but I try to use the most conservative assumptions about order fills.
Our stop is at 33.08
Our limit (exit target) is at 32.04
We were up about $24 on the trade as of the close today.
That sounds very similar to "pivot points". Maybe Williams was the one who originally proposed that idea. BPA could test all kinds of variations on that formula. That's another good item to add to the Pending Research List in the Market Lab. When I get the Geek Academy finished I'll have to go back through these posts to remember everything, lol. At least I won't be at a loss for research projects in the Members Area though! :0)
One of the worst things about those "fat finger trades" on pennies is that they totally blow up the chart! One minute you've got a nice looking normal chart, and then next you have a completely flat chart with one huge spike sticking up out of it, lol. I used to hate that when I traded pennies.
predict todays high and low formula, Larry Williams - you should really pick up his earlier books....
To arrive at the projected high and low for tomorrow, simply add together today high, low, and close. Divide this number by three and then muyltiply that by two.
Finally, subtract todays low and presto - tomorrows projected high appears. To calculate the low of tomorrow, just change the last step by subtracting todays high. The projected low may be a decent place to exit you short sale.......GLTY!!! penn.
Thanks Capitalist. When dealing with pennies though you have abnomalies which mostly are the result of fat finger trades (wrong decimal placement). Not all fat finger trades ar mistakes,sneakin, theavin MM's.
These must be eliminated for a clearer picture, and will louse up your parameters. I have found it most useful by eliminating these abnomalies, by establishing parameters for a "true high" or "true low". Such as high level pps has achieved for two consequtive days (within reason).
When you add in a calculation for accumulation and distribution, certain and apply them to volume (relationship)and daily price action, a pattern will develop, and it is the deviations in the patterns I try to spot.
Ideally you are coming off a short term low preceded by short term high, preceded by a LOng term low,out of a narrow trading channel that is with the bottom quarter of a yealy low. This is the entry position I strive for, and provided I have picked a real company and not a scam, I enjoy an above normal success rate. It is the time factor (patience)that I have problems with LOL.
First identifyingLong term and short term lows I consider essential. A short term high is any high with two lower highes on either side. A long term high has a short term high on either side. These long and short term high patterns are consistant and pretty much as reliable as daily ranges moving big to small and small to big. You know when the range tightens, bigger moves are a comin. The tools above help determine the probability of direction.
I hope you todays ramble interesting. Penn.
Followers
|
4
|
Posters
|
|
Posts (Today)
|
0
|
Posts (Total)
|
187
|
Created
|
02/27/06
|
Type
|
Free
|
Moderator Capitalist | |||
Assistants |
Volume | |
Day Range: | |
Bid Price | |
Ask Price | |
Last Trade Time: |