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.
Myst replied The "scaling" of the deviation bands is a direct result of the movement of price. I'm not sure I follow your meaning of "absolute" and "relative" pertaining to volatility. Perhaps you could clarify? Are you looking more for a Bollinger equation to quantify volatility?
Relative (or percentage, if you wish) and absolute related to price. A relative change has a functional relationship (usually linear) to another number; for example, a price of $40 may have a deviation band at $44. The price changes to $20, and a relatively-scaled deviation band would move to $22. If the change were absolute, the price would change to $20 and the new deviation band would be at $24.
Mostly I wonder how strongly the XDev deviation bands scale with price. I guess this is more a question about whether the market regards volatility in terms of dollars (absolute) or percentages (relative). My feeling is that people (the market) are biased toward looking at dollars instead of percentages, since the lower priced stocks seem more volatile than larger stocks. I also expect that other factors contribute to higher volatility of lower priced stock: less confidence by the holder , volume, institutional support, etc.
Hope this doesn't go too far afield.
best regards Tim
Myst, I looked at this XDev chart of Atari
and I noticed that the deviation lines do not scale much with price. It seems that XDev is looking more for absolute volatility rather than relative volatility (scaling with price). Have you tested scaling the deviation bands linearly with price? Is there some reasoning behind one type of scaling or the other? Is there any evidence that the volatility would remain the same (in absolute terms) and not be linked to price?
tia Tim
It's a G4 Cube.
Here's a page that describes it: http://www.lowendmac.com/ppc/cube.html
V. cool appearance, silent (no fan); too bad it's been discontinued.
best regards Tim
Virtual PC is one option.
I ran VPC on my 500MHz Cube, and tried the XDEV demo. Too slow compared to my PC, but it may be ok on a newer (faster) system. Wasn't painfully slow... might be ok with a 2 or 3x speedup.
Thanks to Myst for posting this interesting, timely and well-written article. Of particular interest:
"...these are debt rather than equity driven economies. The major source of financing does not come from sale of shares in businesses, but from direct loans. There are two reasons for this. The legal structure of Asian corporations gives limited rights and protections to shareholders, who do not collectively control corporate boards. Therefore, maximizing shareholder value is not a driving consideration. It also means that a core measure of economic performance -- the rate of return on capital -- is not a critical variable."
Hello Myst - thanks for your help a few weeks ago with X-Dev 3.1. I have downloaded 3.58 and it's now working well. I noticed that a long optimization period will give no MRI reading - This must be a graphical thing, where the MRI histogram becomes less than a pixel in width. You can expand the window and the histogram will appear.
You posted in message #7223 "you can quickly see the risk of entry here is high. That may or may not mean an entry here will be a losing proposition. It simply means that much of the current short term run has possibly been had and a correction or consolidation of some sort may be in the cards." I know that high is a relative thing, but risk has decreased from 8 to - 5 in one day; can't the "mild buy" be taken literally? I realize that the COMPX price and MRI is in a down trend, and prices and MRI has been lower for longer in the recent past. To me, the short term run ended around 4/13 and the correction or consolidation is underway. Am I missing something here?
tia for your help (and generous atttude about upgrades!)
Tim Reese
MSN screener?
Looks like uSoft has put a bunch of .NET garbage as a gateway to the screener. Went looking for another screener and found the new Yahoo screener http://screener.finance.yahoo.com/newscreener.html
Anybody used it?
Myst, great reply, thanks. Hope that my previous reply wasn't too dry.
I repeat that I think X_DEV is excellent work so far, and a bargain at $128. Searching for the returns optimum automatically would be a useful feature, especially if the automatic method finds the optimum faster than I can manually. Most users would agree that the absolute maximum isn't always the right choice of parameters - but once you have an algorithmic way to find the optimum, that opens the door to lots of experimenting with the "optimality metric" ... you can start incorporating some of those heuristics that depend on up trend, down trend, etc. I'm sure this will be one of the most popular features of V4 <grin>.
BTW I'm now in my 1-month trial for StockWerld Gold (comps of Jimbob closing out my VA subscription). There's lots of stuff there - I've looked at the X_DEV picks, Res/Support and so forth. What would you like to point out as particularly useful and important? BTW the COMPX button (bullish) goes to NASD - not the same are they?
cheers Tim
Revised copy -
Hey Myst - thanks for your thoughtful reply to my somewhat unfocused post.
I guess my use of "alchemy" was intended to connote a more negative meaning than you assign it - more like antiscientific or luddite than anything else. Certainly I agree with you that logic and intuition will serve the trader well. I also feel that there's nothing magic or mysterious about the process - these subconscious algorithms can be put into more concrete form. Development of X_DEV was such a process. I'm sure that the analysis, mathematical or logical, was wholly conscious - it's deciding the relative importance of each input or idea that's subconscious.
If I had to judge (and far be it from me to present myself as an expert), I'd say the main talent you exhibited in development of X_DEV was simply recognizing what works generally. You must have used your systematic, conscious methods as well as those subconscious ones. Finding what works generally is no mean feat - you can always pick indicators (in hindsight) that "predict" a particular event, but which fail in prospective use. This generality poses too hard a problem: X_DEV compromises by performing well only with specific stocks and expecting some external optimization by the user, mostly in reaction to market events. X_DEV's "limited generality" requires that the user pick stocks with proper volatility (often found by backtesting), be aware of and understand market conditions, and have some technical understanding of what happens when you change the program inputs (multiples etc). This compromise seems workable for most users, with a range of success almost always being better than buy-and-hold, and often a lot better. There's still a human in the loop, but X_DEV seems to help a lot, at least IME.
On a somewhat different vector, you wrote
X_DEV's algorithm objectively evaluates price behavior and determines when a buy or sell should take place and how much money should be used based on each individuals particular situation (ie. commitment to the play) and personal trading style.
The interesting points here are commitment to the play and personal trading style. I'd interpret the first as either confidence in the company or willingness to wait out a down trend. Trading style could be a lot of things, but I think you mean the length of the plays that the trader prefers. Some expansion would be welcome.
best regards Tim
Hey Myst - thanks for your thoughtful reply to my somewhat unfocused post.
I guess my use of "alchemy" was intended to connote a more negative meaning than you assign it - more like antiscientific or luddite than anything else. Certainly I agree with you that logic and intuition will serve the trader well. I also feel that there's nothing magic or mysterious about the process - these subconscious algorithms can be put into more concrete form. Development of X_DEV was such a process. I'm sure that the analysis, mathematical or logical, was wholly conscious - it's deciding the relative importance of each input or idea that's subconscious.
If I had to judge (and far be it from me to present myself as an expert), I'd say the main talent you exhibited in development of X_DEV was simply recognizing what works generally. You must have used your systematic, conscious methods as well as those subconscious ones. Finding what works generally is no mean feat - you can always pick indicators (in hindsight) that "predict" a particular event. The quest for generality fits in with your stated preference for comparatively simple metrics; complicated metrics get too specific. X_DEV's "limited generality" requires that the user pick stocks with proper volatility (often found by backtesting), be aware of market conditions, and have some technical understanding of what happens when you change the program inputs (multiples etc). There's still a human in the loop, but X_DEV seems to help a lot, at least IME.
On a somewhat different vector, you wrote
X_DEV's algorithm objectively evaluates price behavior and determines when a buy or sell should take place and how much money should be used based on each individuals particular situation (ie. commitment to the play) and personal trading style.
The interesting points here are commitment to the play and personal trading style. I'd interpret the first as either confidence in the company or willingness to wait out a down trend. Trading style could be a lot of things, but I think you mean the length of the plays that the trader prefers. Some expansion would be welcome.
best regards Tim
Just back from nearly a month away ... regards to all on the X_DEV board.
Just read your interview Myst... not sure if I'd rather picture you as "Myst" or "Russ" - as Russ, you lose a bit of that "man behind the curtain" characteristic that was part of my previous mental picture of you. Learning your level of experience, and that you trade stocks for a living, certainly gives my mental caricature of you more authority.
The most interesting part was your relying on "subconscious evaluation" of all information. This raises two points; first it seems somewhat contrary to X_DEV's objectives, which to me is more of a mechanical system than a subjective system. As I see X_DEV, it seeks to short-cut part of the usual gaggle of conflicting signals provided by the usual approach to TA. By your description, I'd think that successful trading would arise from tossing in a bunch of indicators and taking a guess, then seeing the result and hoping that the subconcious mind would sort out a winning strategy, ie inspired alchemy.
Next, I used to work with knowledge-based systems (a classic area of AI) that sought to codify expert's subconcious algorithms. One of KBS's stated objectives was to elicit from experts their problem solving expertise, which they generally could not express in a compact or unified way. I could imagine a pair of experts, each trying to codify their partners strategy while practicing their own. Might have an interesting result, although the two experts should be different enough in style so that their styles don't converge in the process.
Thanks for removing a bit of your anonimity.
cheers Tim
Hi guys - I'm almost all in cash these days, except for RMBS and my funds (BRSIX, PEMDX). Lots of selling action! Looking for new buys - still a big fan of INGN.
How have you all been? cheers Tim
Oh, I still have 50 shares of DFIB which my broker couldn't sell at the last upswing.
Hello Afaikimo - not exactly. I screened all 88 stocks with X_DEV. I set my start date to 7/1/02, my starting investment to 100%, and started down the list.
My objective was to look through this list of "strong fundamentals" stocks for those that seem to have good X_DEV action. The only thing I did to remove bias was to fix the evaluation period to those 6 mo.
I rejected all stocks that, after optimization with X_DEV, failed to give at least a 40% return over the period and those that failed to beat the buy-and-hold result by at least 50%.
I was not attempting to validate X_DEV, so I don't know what the return enhancement was over all the stocks. I rejected about 70% of the stocks within the first 2 minutes of testing each. After using X_DEV a while, you can often look at the chart and tell immediately, or within a few parameter shifts, which charts to reject. Fixing the period also speeds up the process.
All the charts are posted on VestorAnalytics, or I could post them here.
hth Tim
Paul, you wrote I'm looking for something that will drop, then *>POP!<*. Generally I'd expect to stay with a position for 3-12 months, so maybe my definition of a pop is different from yours. For a shorter term, I'd say that you either have to have inside information or play the technical game to capture a pop.
I believe that, once the current uncertainty is resolved, and once they leave bonds, investors will first be most interested in equities with strong fundamentals. I looked at all the S&P 5-star rated stocks with X_DEV. My criteria were a fixed 6-mo time frame (July 1 '02 to present), better than 40% appreciation over that time, and better than 50% improvement over B&H using X_DEV, $10K initial investment and X_DEV optimization. Here's my list:
ACS
APPB
CMCSA
FCN
FLEX
GSF
MCHP
NBR
PFCB
WEBX
WFT
11 of 88 stocks listed. The best was WEBX, with 80% gain vs. 6.5% loss. I can post the X_DEV charts if anyone is interested.
hth Tim
From the WSJ, 3 March 2003:
Turkey's stock market plummeted by 11.6% today, while the Turkish lira, the least valuable unit of currency in the world, fell to 1.67 million to the dollar, "before the Central Bank announced it would intervene in the exchange rate if necessary, bringing the lira back up to 1,654,000," the Associated Press reports. Interest rates, meanwhile, rose "only by 5.75 percent to 60 percent".
My emphasis. Definitely not an entrepreneurial economy.
Best Regards, Tim
JimBob, I replied to your mail a few minutes ago.
JimBob raised an interesting question on vestorAnalytics regarding the minimum necessary volume for trading. I am currently trading DFIB (57K shares/day), and put up AELA (6K) and MHLX (1.5K) on the vestorAnalytics site. Anybody else have a their own rules about minimum daily volume allowed? The X_DEV manual says to pick stocks with "adequate" volume - I'd expect there to be a tradeoff between the possible high return on these low volume, low priced stocks and the chance that you won't be able to make the trades. Someone must have some experience with these thinly traded stocks.
Cheers, TIm
Myst sed "Anyone else bank some today?"
Uhhuh, NVDA. Would have sold some yesterday but I was offline because of the storm.
Too bad my basis is so high on NVDA ... should have done like Myst and taken a loss on the $13.41 shares. Experience helps, I suppose. Bought more at $10.88 though, and now I'm in the black.
I like NVidia's chances against ATI. A big war chest certainly helps. Anybody else listen to the conference call? Very upbeat.
Cheers, TIm (from his GEForce V8460 equipped PC).
Pretty quiet ...
Here's a couple of interesting X_DEV picks:
Annoying that Holders are so expensive (minimum "trading" block is 100 shares) ... definitely better X_DEV action than IBB or the biotech funds. For the cost of one share of BRK.A, you can trade BBH.
http://www.microhelix.com ... DD needed here.
Cheers, Tim
Well, if we are allowing OT posts....
Myst wrote France, Germany, and especially Russia and China are not concerned with peace......they are concerned about an American presence in the Mid East. PERIOD.
If you'll excuse the metaphor, I wouldn't paint the group's interests with such a broad brush. Within remembered history, Germany invaded the Sudetenland in a cause of liberation with which most Germans agreed. We all know how that turned out. I'd suggest that it's proper, and maybe even reassuring, that modern Germany has such a strong pacifist tendency.
Not to deny the apparently crass economic and political motivations that may bolster the pacifist feelings in Germany, and more overtly motivate the other countries.
Back to your regularly scheduled program ... good call on NVidia, Myst.
Cheers, TIm
Hi CM - re missed limit orders,
Recall that your broker will fill limit orders as he is able in the order received. So, if you have a low volume stock that may trigger a buy or sell, the sooner you place the limit order, the better. Also, there's no reason why you can't let the current day (with its missed order) lap over into the next day. On the day following the missed order, place a one-day sell limit order for 30 sh at $4.87 and another for 40 sh at $4.95 in that order. This would preserve the "works" of X_DEV, as a remedy to the missed sale.
hth, Tim
Wasn't it you in INDX, Lou? Must have made some money today.
Cheers, Tim
Cool ... Elan was damaged goods a few months ago. They're an Irish company, and don't conform to GAAP. Following their mini-Enron scandal, noone would buy because there was no straightforward way to evaluate their PE, though they seem to have been cheap in hindsight. ELN recently sold some assets to KG, for maybe too much money; ELN up, KG down. ELN seems to have shown it can clean house and that it has some good IP and assets.
Here's a few that are in buying range ... usual disclaimer about not recommending them, although they are all of interest to me.
AEterna Laboratories http://www.aeterna.com
Aphton http://www.aphton.com
Millennium Pharmaceuticals http://www.mlnm.com
In transition to a "big pharma" company. A value play, with $4.50 cash per share, though 2003 may be dicey.
Nanogen http://www.nanogen.com
Molecular genetic testing
Novogen http://www.novogen.com
Counters epidermal degeneration (wrinkles).
Cheers, Tim
Zoran (ZRAN:NASDAQ) was a former X_DEV pick ...
CBS MW:
Zoran sank more than 18 percent after the company was downgraded by SG Cowen to "market perform" from "outperform" after the company reported "solid" fourth quarter results but indicated gross margins would decline in the sequential first quarter to around 30 percent from 45 percent. The maker of chips used in DVDs and digital camera also expects to earn 80 to 85 cents a share in 2003, below the average analyst estimate compiled by Thomson First Call of 90 cents.
Seems a weak "badness" to drop 20% in a day. Anybody know anything about this company?
Hey Jimbob, regard IPXL ... generic Claritin approved!
regards, TIm
Yes Lou, Rambus is an interesting story. I've been following the story for a couple of weeks, but hadn't seriously considered investing. Seems that chances are now good for a big runup, but there could be some more good and bad news along the way. One poster suggested that there might be a good entry point at the Iraq invasion... might be worth watching for.
Cheers, TIm
In honor of the POTUS, something conservative:
Cheers, Tim
NVDA - Negative coverage?
Market Edge Opinion - "AVOID: If you are long, close position or monitor stock closely.
Moving Average Convergence/Divergence (MACD) indicates a Bearish Trend. Chart pattern indicates a Weak Downward Trend. Relative Strength is Bearish. Up/Down volume pattern indicates that the stock is under Distribution. The 50 day Moving Average is rising which is Bullish. The 200 day Moving Average is falling which is Bearish."
Don't know that I agree, but this might explain the market sentiment. Personally I think it's a good buy now ... I own some.
hth, TIm
An interesting story - worth watching if you dig this sector.
PR News Wire - Introgen Therapeutics, Inc. (Nasdaq: INGN) announced the publication of data from its phase 2 study combining ADVEXIN, an adenoviral vector containing the p53 tumor-suppressor gene, with radiation therapy. The study, treating patients with non-small cell lung cancer (NSCLC), showed that approximately 60% of patients' primary tumors regressed or disappeared after the combination therapy, as assessed by both biopsies and by CT scans three months after treatment. Moreover, ADVEXIN administration did not appear to increase the side effects caused by radiation treatment. These data were published in the January 2003 issue of the journal Clinical Cancer Research. (My emphasis)
Cheers, Tim
Yay - 1st X_DEV profit.
DFIB, up 13.3% since December, about 4.5% realized.
Making money sure is better than not! <gg>
Cheers, TIm
hi, you're welcome.
Anybody notice BINX? I posted it last night on the VestorAnalytics site. Lots of gaps, up and down - I was able to backtest 262% over 3mo. Too bad (for me) they announced they're being purchased this morning for almost a 50% premium. Congratulations to the shareholders.
Cheers, Tim
Hey - hmmm, tell me if my interpretation is correct. NBI is orange so it's moving sideways. NBI is most highly correlated with small cap and semiconductors, and least correllated with the broader market (DJ, NYSE, S&P 500). I guess the shape of the tree isn't important - it just fits in the available space.
Cheers, Tim
Yeah, I put a few more on the VestorAnalytics site. The added 5 are biomedical equipment and supply companies. Not as pretty as Myst's picks, but they flop around and (potentially) make money.
Those who haven't been to VestorAnalytics, take advantage of JimBob's free peek period and take a look...
Cheers, Tim
Hi Lou - "I ... was interested in the ones that seem to have more cash in their coffers for a longer ability to operate. (cash/cash burn)."
Yes, none of these companies are profitable, so the first cut in estimating value usually depends on a prediction of the company's ability to become profitable before going bankrupt. Typical inputs would be cash burn, cash reserves, growth of revenues, and product pipeline. The risk is high that the company's potential won't be realized, so the usual advice is to diversify between your best picks - and maybe some of your picks will hit big and make up for the rest. There's still a lot of negativity in this sector because of recent history... especially wrt companies that aren't profitable.
I'll follow up with some more specifics about companies later.
Cheers, TIm
Low-priced, high-beta biotechs, part 2
In order of highest beta. Burn estimates are from Q3 reports.
Genzyme Molecular Oncology, beta 1.78. Technology and Human Healthcare company that develops innovative products and provides services for unmet medical needs. The Molecular Oncology division develops cancer products focused on cancer vaccines and angiogenesis inhibitors. 79.6% 1yr revenue growth, 5.6M debt, 1.60 book, 18M cash, 6.3M/q burn.
Corixa Corporation beta 1.71. A developer of immunotherapies with a commitment to treating and preventing autoimmune diseases, cancer and infectious diseases by understanding and directing the immune system. 16.3% 1yr revenue growth, 16.3M debt, 6.78 book, 116M cash, 23M/q burn.
La Jolla Pharmaceutical Company beta 1.59. Biopharmaceutical company focused on the research and development of highly specific therapeutics for the treatment of certain life-threatening antibody-mediated diseases. 0% 1yr revenue growth, 1M debt, 1.37 book, 65M cash, 11M/q burn.
Immune Response Corporation beta 1.55. Biopharmaceutical company developing immune-based therapies to induce specific T cell responses for the treatment of HIV, autoimmune diseases and cancer. Also developing a delivery technology for gene therapy via intravenous injection. -99% 1yr revenue growth, 7.7M debt, 1.37 book, 0M cash, 11M/q burn.
XOMA Ltd. beta 1.49, Biopharmaceutical company that develops and manufactures recombinant antibodies and other protein products to treat cancer, immunological and inflammatory disorders, and infectious diseases. 96% 1yr revenue growth, 66M debt, 0.19 book, 36M cash, 12M/q burn
I'm not recommending any of these per se; they're just interesting X_DEV canddates. The fundamentals listed should give you some clues for what to look for in addition to X_DEV return.
Pretty quiet ... here's some grist for the mill
Low-priced, high-beta biotechs, part 1
In order of highest beta. Burn estimates are from Q3 reports.
Orchid BioSciences, Inc., beta 3.64; Engaged in the development and commercialization of technologies products and services designed to measure and use information related to genetic diversity. Applications include health care forensics paternity testing and improved crop development. 1yr revenue growth 141%, debt 7.1M, book 2.02, 12.5M cash, 8.4M/q burn.
Illumina, Inc., beta 3.23; A developer of next-generation tools for the large-scale analysis of genetic variation and function. 1yr revenue growth 214%, debt 26.2M, book 3.31, 74M cash, 7.6M/q burn.
Sequenom, Inc., beta 3.11; A discovery genetics company that has integrated a technology platform, sample repository, assay portfolio and strategies for determining the medical impact of genes and genetic variations, known as single nucleotide polymorphisms, or SNPs. 1yr revenue growth 35.5%, debt 10.8M, book 8.23, 106M cash, 17M/q burn. Look at Q4 outlook.
Abgenix, Inc., beta 2.51; Biopharmaceutical company developing and intends to commercialize antibody therapeutic products for the treatment of a variety of disease conditions. Includes transplant-related disease, inflammatory and autoimmune disorders, and cancer. 1yr revenue growth 11.6%, debt 200M, book 9.11, 464M cash, 34M/q burn.
AEterna Laboratories, Inc., beta 2.12; Biopharmaceutical company focused on the treatment of cancer. Its lead product, Neovastat®, is the focus of a number of clinical trials including pivotal trials for the treatment of lung and kidney cancer, and multiple myeloma, a type of blood cancer. 1yr revenue growth 257%, debt 9.7M, book 1.50, 89M cash, 6.6M/q burn.
Hope you find this interesting. Comments about the X_DEV setups are particularly welcome.
Cheers, Tim
Lou, whitelake, my best guess -
There's nothing wrong with whitelake's internet feed or setup.
The Yahoo data is fine.
The problem with the SI data is a bug at the SI website; look at the historical data on the website http://www.siliconinvestor.com/research/historical.gsp?s=QQQ . whitelake, try IBB or another ETF and the result should be pretty close to the Yahoo data.
There are two problems with the Wallstreet City data. First, the chart date is always February 8 or 9, regardless of the selected date. This is a bug in the release version of X_DEV, and has been fixed in Lou's beta version. Set the selected start date to 12/5/01, and the chart will range back to 2/9/01. Second, the values are different from the Yahoo or SI data. This is something different about the WC site; that's why I asked about dividend rescaleing previously.
best regards, Tim
Dunno ... Wallstreet City data has seemed goofy before. The data range doesn't match the selected range, and the data somewhere drifts away from that given by Yahoo and SI. I'd guess partly a feature of WC and partly a bug. Does WC rescale for dividends? The data for Yahoo and SI is the same, except that SI is listing the high for the open. This expands the scale of the chart very slightly for SI, and of course, screws up the deviation bands. I'd just use Yahoo.
hth, TIm
Hi Charlie - I think that you are right if you pick the right stock. You can certainly pick stocks that, with such conservative settings, will have lower returns with X_DEV compared to buy-and-hold (I'll have to look for an example). Some charts fit X_DEV to a tee, some don't. I'd guess that Myst was interested in particular stocks (ie semiconductors and such) and that X_DEV was developed to fit his preferences.
IMO you can be sytematic about suboptimal settings - and I also believe that you are likely to have less risk at a suboptimal setting. For example, pick one of the better performing X_DEV stocks, optimize fully, and start backing off toward tighter bands and lower multiples. Generally you reach some break point up to which the returns didn't decrease much, but beyond that they start to fall quickly. I'd guess that at that break point is the most likely place to get the best returns for the least risk.
JMO though - I've only been X_DEVing for a few months, and haven't made money yet (mostly because I'm more interested in biotech and genomics than the semiconductor stocks that have been doing well).
hth (hope this helps), Tim
Anyone else have an opinion?
Opinion? Sure! <gg>
Other than just looking at the bands and guessing what the prices will do, as Steve suggests, I don't think there's much hope for consistent X_DEV results at less than multiples of the 12-day SMA period. Certainly at fewer days than that, short term cycles will be filtered strongly, to the extent that 3, 4, 6 and 12 day cycles are completely removed. Maybe Myst can, but I can't really predict by inspection what results the deviation factor times the buy&sell multiples will give... in most cases I have to actually run the program and see. You can use the deviation bands to eyeball the prices, but it's not really X_DEVing.
Plus, for the backtesting, there has to be some optimum that trades off enough samples to be significant (ie going back far enough in time) and not including so much history in the backtest that it becomes irrelevant to what's happening now. If there's some real predictive power to X_DEV, then it must be filtering out the information that's important at its time scale, which I've always thought was between 30 and 180 days, give or take a few.
JMO. Cheers, Tim
Caspar, pick iHub FAQ at the bottom of this page, then select Posting Features for an explanation. If that isn't enough, post again and I'd be happy to help.
I wondered about Rambus - what kind of return does your setup give? Didn't they have some kind of accounting scandal recently?
Cheers, Tim
"Thanks for the feedback on placing limit orders early. Frankly, it never even occurred to me."
Hey Lou, you're welcome. I discovered this aspect trying to place a limit order for DFIB. My broker offers an "all or none" checkbox, which I was picking - seemed to make sense since I didn't want a partial trade of a low-priced stock. Turns out that checking "all or none" made my trade restricted, and restricted trades go to the end of the line. This went on for several days; 3 or 4 times the prices briefly dipped into my X_DEV buy range, but no purchase - there were just too many trades in line ahead of me. WIth a little research into the fine print, I discovered the restriction. Once I stopped picking "all or none", my trade went through, although in two parts on the same day.
Too bad DFIB's been going sideways since then.
best regards, TIm