| jfcampbell3 |
|
Gold User, Member, Platinum User, TeleChart
|
| Registered User |
|
|
|
|
| Unsure |
|
| Friday, November 30, 2007 |
| Sunday, August 30, 2009 7:43:16 PM |
17
[0.01% of all post / 0.00 posts per day] |
|
|
As the CTO (aka chief nerd) for a software development company, I am always working with management as well as the development staff regarding software upgrades to our product lines. There are a couple of things I'd like to comment on, given the earlier posts to this forum.
1. The non-committal response (as to if / when a new upgrade will be forthcoming) indicates to me that there are no plans. If there were plans, the word "if" would not be used by a staff member in public. Also, "when" means there are no plans, because there is always a struggle between management, who set unrealistic time goals, and development personnel, who hate being pressured. If there are plans, there is always a "when".
2. I also understand a couple of other things about these issues, which allow me to be a bit less critical:
a. My guess is that the bulk (all?) of the development staff is involved in the Blocks effort. (From my own perspective, it was a pretty good effort, but not good enough to get me to use it, even though I invested quite a bit of time and money getting to a training and evaluating it rather thoroughly.)
b. My other observation - again completely guesswork - is that the underlying technology of TeleChart is not very amenable to modular upgrades. Yes, bits and pieces such as adding new indicators, etc. would be feasible, but replacing some one piece like the watchlist manager may not be possible. If my hunch is correct, they'd have to rewrite the whole architecture to easily change any one piece, or add significant new functionality. If you're going to do that, you might as well build an entirely new product. This is a long, expensive, and risky thing to do. It is also problematic since Blocks is an entirely new product, and multiple major product design, implementation and roll-outs have lots of fiscal, marketing, management and sales implications which almost always drive technological decisions.
c. If my two hunches (a and b) are correct, then Blocks has to generate enough revenue to warrant such a rewrite, which also takes a long time.
In the end, I find myself torn: the technology is rather good, if archaic; I find the forum very useful, support and training is good; and I still have a lot to learn from this environment. But I would concur that the software needs updating in order to retain the customer base, or it will soon be overtaken by other packages on the marketplace. And if Blocks doesn't prove to be a block-buster  , then there are other problems bearing on investing in TeleChart.
If I were in charge, I'd see if I couldn't roll out a few features here and there to keep people happy, and figure out how to modernize the thing, if at all possible.
Just 2c worth from someone who has a lot of empathy for the complications involved.
|
Diceman, Stockguy, et al.
Thanks for your comments.
However, sorry to say, I'm just not getting this.
If I have a 4-bar / 8-bar moving average comparison, at the point the 4-bar crosses the 8-bar, the MACD histogram would be 0, right? If I plot a 4-8-4 MACD histogram this does not happen. (All indicators are exponential).
Re: stockguy's input, I get the point that the leading 4-8 portion must correspond to my two MA's but it appears to me that the 3rd figure -- the period -- is the crucial one here, and in poking around on the web for MACD formulas, I didn't come across a good explanation of the meaning of the periodicity.
Continued thanks
John
|
I am working with moving average pairs, testing different values for crossover indicators. Examples are 3 bar / 7 bar or 4-bar / 8-bar. I would like to compare a MACD for these to see if it might be a leading indicator, but am not sure how to structure it. For example, if I am looking at a 4-bar / 8-bar moving average, and try a MACD of 4-8-10 it doesn't seem to indicate (visually) the difference between the 4-bar and 8-bar moving average.
Sorry, I still am not adept enough to copy and paste a chart into this window to show what I mean. But if the 4-bar MA is moving at an accelerating rate to the 8-bar, I'd like a MACD that would indicate this.
Also, in the short / medium term (<1 day to 5 days) does anyone have recommended pairs of MAs which they have found to be useful indicators?
Thanks
John
|
This is also posted in the suggestions forum, but I thought it might get a bit more viewer hits here (so far, 0 over there), to get feedback...
One of the areas I find in great need of improvement is the handling and management of watchlists. It appears that the same software algorithms have carried over to Blocks, as well, so I hope that others share my feelings and will add to my observations. (I do like the feature that TC watchlists show up in Blocks with no effort).
First and foremost, I feel that watchlist management needs to be hierarchical rather than the flat system currently employed. Everyone knows what this is: the windows explorer is just that. Thus, you should be able to group watchlists within categories: December 07, My tests, etc. etc. This, you could start with just a list of the watchlist groups, (rather than all of the items in all of the groups being visible all of the time).
If you want to move a list from one to another: drag and drop; or when you are done with a group, delete the "folder" containing that group. (Deleting a bunch of old watchlists is really painful). And so on.
The other thing about hierarchical organization is that one can drill down rather quickly rather than having to hunt through a flat list. This is also helpful since there are, for example, on my installation, numerous system lists which have no entries, and others I'm just plain not interested in, but in a flat system they're there and I have to wade through them.
Also, I don't understand the technical reason why a stock cannot be removed from a list generated by an easy scan. Even if there is a valid reason, it should not continue to be true if I flag stocks from an easy scan-generated list and copy them to another (or a new) watchlist.
Well, there's more...I guess I use watchlists a lot, and find it really wanting a lot of sprucing up.
Watchlists should have an option wherein all the UNflagged items can be removed. If I flag a whole bunch of stocks as worth watching, I shouldn't have to copy them to yet another watchlist in order to get rid of the ones I DON'T want to watch.
Also, one should be able to cursor up and down in a watchlist. I should be able to use the up arrow to move back up the list if I choose; and in addition to the space bar, I should be able to use the down-arrow key to move down a list (and, in my opinion, page-up and page-down).
In addition, since at the watchlist level the type-and-filter option works, it should work for the stocks within a watchlist, rather than having to "jump" to get to a specific stock.
If I type in a part of a name at either the watchlist top-level or within a watchlist, while I do like the "filter" option, I should be able to down-arrow into the filter list to select the watchlist (or stock) I want to select. (In compliance with windows and mac standards).
While I'm in the arena, there is another thing which I would like to see changed: you cannot set some parameter to see only stocks with notes in a specified date range. Frankly, I'm not interested (99 44/100 of the time) in notes that are several years old. I would like to say "show me stocks with notes from the last 30 days, or last 10 days or ..."
I'm sure there are scraps of paper floating around my office with more things I'd like to see done to the watchlist portion of the software, (I have quite a bit written down on the other sections, but they'll have to wait until another day) and I'm sure many others have their own insights. Since this entire mechanism is currently being carried forward into Blocks, as I mentioned, I would like to see (if others agree with me) a strong lobby to improve this vital part of a great pair of products.
Cheers
John
|
I posted a question under the thread "sir rocket scientist's indicators" along the same lines as others here. It does not make sense to me that both formulas ended in a truncated boolean, so rather than repeating that whole post from the other thread, I'd ask, ff you will, take a moment to read my post there.
I'd really like to see this resolved. I don't believe that the TSV3 > 0 posulate explains the fact that Oversold's doesn't have a boolean comparator either. Plus, as I said in the other post, I'm too new to understand the thing, anyway.
Can we get Dick to answer us if the community or a trainer doesn't have an answer?
Thanks
John
|
As a newcomer to TC, I would like to ask clarification of this same formula. (The issue for me, is the same for either formula - oversold or overbought).
Here's the oversold formula.
L<((2.95*MINL15.7)-MaxH7.3)/2 and H>H1 and STOC50.3>50 and TSV3
I would "translate" this something like
Today's low < 2.95 times minimum of low periodic 15 days as of 7 days ago minus the maximum high of 7-day period as of 3 days ago divided by 2, and today's high greater than yesterday's high, and stochastics 50-day period as of 3 days ago > 50 and TSV 3-day period...
OK, so I know I'm still learning to translate TC symbols into English, and I don't know all of the punctuation, nor all of the indicators etc. So I am sure that I am way, way off in what I'm trying to say. My first problem isn't really whether I got the translation right (or even close).
It is that I get lost on what "logical" TSV3 represents. Even if I'm completely off on all of the rest, my days of programming and logic courses tell me that any function on either side of an "and" must resolve to a logical yes or no, and I can't get through my thick skull how TSV3 all by itself resolves to yes or no, much less how it fits into the picture.
I checked the formula and it did resolve without an error (witness the other post) but if this question were on an SAT (equally incomprehensible to me!) I would have failed.
On top of which, quite frankly, I'd like to understand what this guy's formula is trying to tell us, and I don't have the experience to "get it" - as witnessed by my pidgin rendering.
I find this point very interesting, especially as we appear to be heading into some bearish times, but as a novice, it is more than a bit cryptic to me. I get his message, but I'd like to get his math, too. Can a trainer or anyone else help render both of these puzzlements comprehensible to me?
Thanks
John
|
I searched through help looking for a means of comparing price and Bollinger bands. All that came up was quickedit, which I assume is a Blocks feature. Is there any way to create TeleChart PCFs for BB values? (And envelope channels, for that matter). For example, I'd like to create a %True when price is above the upper BB and crosses over (below) it.
Thanks
John
|
Thanks, all for the input and feedback. I will refine the whole theory and post it sometime soon. It may or may not be of value, but it's been a great learning tool for me.
Sorry that the pic didn't come through for some of you. Might be specific to your browser, because I could see it in my (2nd) preview, and can see it in the post today.
Good holidays to all!
jfc
|
Thanks to davidjohnhall for his insightful means of putting a chart in a posting. I do it a bit differently, so I thought I'd share a couple of other means - for Windows users. if you have a chart up in a window, you can copy it to the clipboard with the alt-print screen. Find the ptrsc key (may be labeled slightly differently on different keyboards) hold down alt then prtsc. Then you can copy it to paint as David said, or you can put it into word, bring up the drawing toolbar (view-toolbars) and hunt around for the crop tool. Once you've done that, and have the part of the chart you want, then put it into paint, save it as a bmp image type. Then David's instructions for clicking the "insert / edit image" button and thereafter work fine. It's the same process, but a different way of approaching it.
jfc
|
|
I am trying to understand %-true indicators as a means of validating some trading ideas I'm working on. pls. forgive if this is a double posting - I spent an hour preparing it, and when I clicked on preview, it disappeared and I haven't found it in the current discussion queue; so, here goes (again).
Suppose I have two predictors which are "true" in high correlation - does this, in principle, make one a reasonably good predictor of the other?
Example:
Here, you see that one indicator, the white one, is a pointed-top or teepee 100% of the time. I am using it as my "predictor". The other, the red one, which involves price direction (isn't that what we're all trying to predict?) can be a teepee, but is more often a flat-top i.e. can be true over a short period of time. This is a snap of a daily view, by the way.
If the first (white) predictor becomes true during the time the 2nd is true, and above all, prior to the time the 2nd is no longer true, can it be said to be a short-term predictor of the 2nd?
I will be happy to share and refine / improve my indicators if I have understood things correctly, but there's no point in dragging out a red herring if I'm off the mark. They're not very complicated, and are simply the result of thinking about the few principles I understand, doing a bit of observeration on some charts, and coming up with a couple of formulas to test them out. So far, so good, but I don't want to waste much more time if I'm heading down the wrong path.
Thanks!
jfc
|
|