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.
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.
Eagles!
~~~~ COMPX 12/03/2002 ~~~~~
Previous Close 1484.80
1478 AKvetch
1498 WTMHouston
1505 Phil
1518 timhyma
1522 MM
You must have one of those Canadian calculators! I'm lost again. What do you mean?
~~~~ COMPX 12/03/2002 ~~~~~
Previous Close 1484.76
1513 MM
Congrats AK!
TEST
~~~~ COMPX 12/02/2002 ~~~~~
Previous Close 1478.84
1486 AKvetch
1498 WTMHouston
1509 MM
1512 timhyma
1519 Muell (hehehe)
Congratulations SagDec15!
~~~~ COMPX 12/02/2002 ~~~~~
Previous Close 1478.84
1498 WTMHouston
1512 timhyma
1516 MM
I think you can get it here in txt format:
http://www.traders2traders.com./linkindex/Default.asp?CategoryID=3
The reason he skips the opening bar or half bar is that it takes a little time for the signal to reach an equilibrium or stabilize. The open is apparently too unstable for this type of an indicator. Also he favored natural hour bars that end with the little hand straight up on the clock. Delaying the start until 10AM EST achieves that goal although there are other ways to achieve it and I'm not sure it's just the first 30 minutes of the session that are excluded. For instance TS has a switch for natural hour bars that adjusts to the hourly close for hourly bars even if the initial bar starts on the half hour.
I believe he originally was experimenting with NA for Net Advancing issues. I thought that meant advancing minus declining issues. That might be different than the version he released as oddball but NA is what he called it prior to oddball. I always thought it was the same thing and he just changed the name.
~~~~ COMPX 11/29/2002 ~~~~~
Previous Close 1487.94
1502 MM
Happy Thanksgiving all.
I think Zeev said he used total Nasdaq volume and I think oddball uses net advancing issues of the NYSE?
I think it might work. Sorry I don't have any thoughts to add to it at the moment but your list includes some good one's.
Thanks, MM
Why do you prefer to base your stop on $NDX when it's impossible to place an order using ndx as a stop? I know you can exit manually when it's hit but there must be a pretty good reason for not basing it on QQQ.
He gets a congratulatory message from me! #msg-600304
Congratulations! #msg-600000
BTW, I could easily be mistaken about oddball using specific volume and certainly there's differences with what Zeev does but I'm pretty sure oddball had an unusual way of determining if today's volume would trigger a signal. Best I can remember it was based on a comparison to yesterdays volume. Typically volume is compared to the previous bar rather than the prior days bar. Unless looking at daily bars but both of these specific volume descriptions reference intraday bars so jumping to the prior day's same time is quite unusual but yet it's unique to both oddball and Zeev boxes.
That's the one. And after rethinking things I'm looking at it again because as I recall I believe it uses specific volume? ...that's that unusual volume that Zeev mentioned but in this case it compares today's volume at a given time to yesterdays vol at that same time. Also, it wasn't clear to me if Zeev was using total vol or had broken it into advancing minus declining issues. Anyway it should be obvious I've never fooled around with it and still have a lot of unresolved questions about it.
I'm thinking it might be an adequate filter for another strategy. The first time I looked at it there were too many quirks like it ignores the first and perhaps last half or full hour of the day which Mark determined was unreliable for some unknown reason. Those type of quirks make it difficult to implement but not impossible.
What link? I didn't see it.
It's looking good!
LOL! You must really enjoy that graphics stuff because you're quite good at it.
There isn't any correlation with what we do here and the market. It's a diversion with it's own unique edges that aren't transferable to trading.
LOL! It's that blue sky that's captured whenever we wag at either extreme of the list. I've never studied it but it would be interesting to see how many times the top or bottom spot won. Yesterday SagDec15 won using that strategy. Today there's a big patch of blue sky between Phil and Muell. It's unusual when it occurs in the middle of the list. I had my eye on it hoping Muell would raise again and I would stay put but he didn't go for it.
I've never seen anything that produces a positive expectancy after commissions and slippage posted on the net and I've tried most of the ideas that people have posted. If you do a google search for "oddball" and find a system that uses the volume of advancing and declining issues then that might be an exception. It's free and it's supposed to be profitable. I've never cared for the idea enough to to look into it.
No I don't release things like that because I respect the owner's efforts to develop it and my own in reverse engineering it. With that in mind let me add that it's no big deal. It's not hard to figure out how most things work when I'm motivated. And there's nothing special about my skills. Anyone should be able to do the same with enough practice.
At first glance it makes sense to buy on the expectancy it would rally to close greater than 5%. It would be more prudent to first check the correlation between the few outliers that close down greater than 5% and an intraday value that was below 5%. Obviously all down closes greater than 5% were down that much at some point intraday. The important test would rank those intraday swoons against the ones that still existed at the close, rather than ranking them against the mean which rarely closed down 5%.
That type of design error in determining what to measure something against wasn't included in my mention of bugs. To me a bug is a typo that references the wrong value of a variable. That typically produces unpredictable results going forward, unless the bug itself captures a market characteristic by accident which isn't likely. Bugs are also logic errors that juxtapose the flow of the program in a way that although it's consistent it's still likely to produce unpredictable results.
Design errors where something important wasn't considered fit into some other category than bugs. If I'm determined to do something and a good idea of what it is then I always figure out a way to accomplish it. However the type of correlations you're discussing seem at first glance to be quite different than what I normally do. But prediction based on historical norms and outliers is a cool area to investigate. Good luck with it.
COMPX gave us a downbar today with H<H[1] and L<L[1]. Close confirmed the change from upbars to the first downbar by giving us c<c[1]. If I was using just those measures to determine direction I'd guess tomorrow is another day with L<L[1] and H<H[1]. Since it would be the 2nd downday I wouldn't require c<c[1] to confirm that it was down. However with as strong as the market has been moving upwards if tomorrow meets the lower L and lower H requirement that would define a thrust extension. Under those circumstance if the close reverses up with c>c[1] it implies the thrust extension exhausted itself and turned into a thrust exhaustion. If tomorrow is a lower L and lower H combined with a down close I'd be looking for it to continue down.
I believe some proprietary swing algorithms that are quite popular triggered downswings today. I don't know for certain because I haven't looked at their charts but my knowlege of how it's calculated tells me it's reversed down.
The daily compx of that experimental signal I'm still tweaking reversed down today too.
But even with all that I'm bullish because it's not entrenched enough for me to give up my horns. The weekly compx is an upbar so far and would have to fall considerably to trip L<L[1] and even then wouldn't be a downbar but instead would be outside because it's already exceeded H[1]. It could finish as a close reversal or an outside close reversal?
Darvas describes a shadow that extends 5% beneath the box that he calls the danger zone. I don't recall the exact stop level for when price falls back into that shaded area but I see a problem. 5% isn't adaptive to the timeframe. It's based on an optimization he did on weekly charts. On a 5 min chart 5% is way too much when the entire day's price movement often is contained within 2% movement per day
Zeev on the other hand has defined a stop below the box that adapts to all time frames. The same reference works regardless of whether it's a weekly or a 5min chart. At least that's my understanding of it and Zeev might correct me if I'm wrong.
Regardless, I think it's important to develop these types of variants and welcome ideas along the lines of building a hybrid Darvas box. I don't expect the orignal algorithm to work, particularly because it's been published in a book and widely diseminated.
~~~~ COMPX 11/27/2002 ~~~~~
Previous Close 1444.43
1414 timhyma
1420 Albert
1447 AKvetch
1459 SagDec15
1464.43 Phil
1484 Muell
1488 MM
You're disgusting.
Disgusting.
Describe it and I bet I will remember it and know where it is.
What secret code would that be?
~~~~ COMPX 11/27/2002 ~~~~~
Previous Close 1444.43
1420 Albert
1459 SagDec15
1477 Muell
1481 MM
Nope, give me a clue where I should look.
Yeah I agree that's what I'm saying. But more so ;) I have to get that word absolute into the conversation to be sure I'm agreeing to what I truely believe. And that is absolutes exist for me. Sometimes it's a real bitch because I can't make something work unless I fudge the results or look the other way and I'm unwilling to do that. I have to know exactly how the process proceeds and know without a doubt that there isn't anything in the process I'm not aware of that's influencing the results.
There's obvious things I miss and don't understand but my only concern is that the parts I do understand are doing exactly what I've programmed them to do. Otherwise I have a bug and that's a random element that I can't count on to continue producing the same results. Even though it's consistent and would always give the same results over the same data it likely would be entirely unpredictable on unseen data because the nature of the bug usually doesn't have anything to do with a price characteristic that I'm expecting to continue to repeat. In other words if a bug causes the strategy to give outstanding results it usually can't be counted on to continue.