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.
I think it would be instructional for everyone to set up one or more fictitious (model) portfolios and apply recommended trades within those portfolios while following preset rules. This would allow meaningful comparisons between different approaches.
For a starter in setting rules:
1) have two portfolios
2) one of which allows short trades while the other one does not
3) the no-short fund may go long an existing (not fictitious) fund which shorts
4) each portfolio starts out with the same amount of play money, say $10,000
5) stocks held long cannot be used as margin for stocks shorted -- only cash allowed
6) in the event of a margin call, funds must be added to the portfolio (at an annual cost of prime + 3%) and the funds must be removed when the position is closed
7) cash held as margin earns typical money market rate
8) current bid/ask is used at the time of trade decision
9) a trade fee of $10 is deducted on every equity trade
10) entering and exiting a short fund may only be done at the end of the day and all typical fees must be accounted for
11) post results frequently on the board in a tab delimited format containing named columns and rows
12) each transaction class is accounted for in a unique column of the data table
13) the last row of the data table is a SUM(above) of all data columns
14) the value of the fund would be a SUM(left) of the last row of data
I still don't quite follow the reasoning, even for comparison purposes. For example, two separate trades of equal amounts going long, each of which yields separately 6% do not add up to a total yield of 12% unless those trades were performed sequentially. I do not see in your table the opportunity to execute separate sequential long trades which add up to 13.2%. The long trades overlap.
Regarding going short, a valid comparison still requires taking into account the requirement to hold 50% margin. One short trade does not constitute a portfolio of trades even though there are two holdings (borrowed stock and margin); there still remains the absolute requirement to hold margin. In other words, a short trade of $1,000 is in reality a trade of $2,000 given the requirement to hold margin. Therefore, a gain of 6% on the $1,000 is a true gain of only 3%. This is especially true when adding gains from going short with gains going long.
"Trades Status (BUY1 and SELL1) T%Gain= 13.2%"
BUY1 and SELL1 seems to infer only going long, and I can't get to 13.2% without investing money which is not available because it is already invested.
I don't understand the accounting on the short side either. Total gain would have to take into account funds held as margin, which usually earns very little because it usually sits in a money market account.
Sorry for being a little too vague.
I'm referring to what you call the SMH20, which you use for BUY1 and SELL1. You have been basing, in part, your buy/sell signals from that indicator when it bobbles around +100 or -100. Since SMH20 only contains 20 stocks, it will be sensitive to movements in a single stock. My suggestion is an attempt to take some of the noise out as an individual stock may have, because of a specific news release related only to it, a significant effect on the overall outcome.
Les's analysis wont have that problem because of the large number of stocks in his calculations -- over 6 times as many.
Hope this is clearer.
Comments:
In statistical analysis, in order to obtain a more representative indication of a trend, it is sometimes useful to temporarily eliminate the few data points which vary significantly from the norm. In this case: AUGT, VECO, and AMKR come to mind. Since you draw such a fine line (less than 1%) between sell/buy (or not to sell/buy), I wonder if temporarily eliminating standouts would give better indications.
From what I understand from previous posts, Les's formulas are considered proprietary. I think there is a recent post where Larry said as much.
You might try an internet search on those names, then play with modifications of the formulas you find.
Similar (and well known) things I know of from MetaStock are Chande's Trendscore and also a phase number derived from a modified version of the (somewhat) standard MACD (but using different delays, and perhaps even using a mathematical derivative of a function because a flattening of a curve can signal a trend change and a derivative of that function would give you +/- indicators of various magnitudes as a result).
My elementary understanding of the analysis done on this board is that only price movement is analyzed and there is no consideration of volume. If you want to do some work, I think it is very worthwhile for you to consider how to incorporate volume in your indicators. I think that approach (i.e., including volume) would make it easier to more accurately predict intermediate trend changes.
"132 gains chart indicates this rally has no depth. IMHO."
How did you count gains(losses) of 0%? Without doing the count myself, I suspect you counted those companies as losses.
Another way to interpret the data is that the gains, though without depth and strength, have legs -- if a much higher percentage of stocks had sprinted forward, then one could justifiably expect a correction.
"The potential for an inverted yield curve is growing."
On a percentage basis (and certainly on an absolute basis), when comparing current rate-inversion possibilities with past possibilities, the yield curve was much closer to inverting during the period between the end of 1996 and the end of 1999. I wouldn’t have wanted to miss that stock market move because of fear of rate inversion.
"It's like you're arguing that the FED raising rates is a good thing? It's not."
Before the recent rate increases, the short rate was so far below par that for all practical purposes it was essentially at zero. That condition, though it benefited the economy at the time, gave the FED no flexibility on the down-side unless they wanted to start "giving" money away in the manner that Japan has had to in order to get their country out of their multi-year (near multi-decade) recession. The FED is now slowly raising short rates while we have a good economy with the goal of achieving par. That would put them back at the point where they would have rate flexibility (meaning, the ability to either raise OR lower rates depending on the economic situation). To that end, raising rates IS very good.
The table confuses ... why are the two trades with entry dates of 2/8 (9:00 and 8:00) at the top of your table showing positive gains? Obviously, I must be misinterpreting your table. I think you are trying to point out that there would have been aditional gains if the original position (entered into on some much-earlier date) was sold on 2/8 rather than on 2/9, in which case you should change the text in column 1 to better reflect what you are trying to say. It seems that you are using the table to say various (and different) things even though the entries in column-1 are always the same.
"... What happens historically when the FED raises rates six times in a row?"
How much of that interest rate data was at a time when short rates were at or above par? Not having researched it, I would still guess virtually all of it.
This time, things are different. After six consecutive rate hikes, short rates are still well below par. It will take several more rate increases to achieve par. This time, any sell-off of equities as a sole result of [unwarranted] rate-hike fears may produce very good buying opportunities. The phrase "sole result" is key, as there are other reasons why the market can sell off.
Thanks. It finally dawned on me after staring at the actual numbers that the second table is not anything more than a lookup table that you say the Feds use.
I'm with ajtj99 ... your explanations need a whole lot of work. I agree with your conclusion regarding probability or recession, but I don't know what the heck your second table shows.
I understand the first table, in which are listed the symbols, period, range of values (three columns), and spread (last column). The other, longer table, doesn't make any sense.
Questions regarding that table:
1) Is there any significance to the numbers listed under "num"? I assume they are Excel row numbers, and have no significance.
2) What is the time period between the different rows?
2) What is the label for the second column?
3) What are the numbers in the second column? 10-year spreads? None of them seem to reflect any part of the numbers in the first table.
4) How do you arrive at the number 100 at the bottom of the second column? Arbitrary? Same units of measure as the numbers above it?
5) What are the numbers in the third column? 30-year spreads? None of them seem to reflect any part of the numbers in the first table.
6) How do you arrive at the number 200 at the top of the third column?
7) How do you take the numbers 100 and 0.76 and come up with the number range of "5 -10" in the fourth column?
Starting a Dogs of the Dow portfolio?
"Timer Digest" has Tim Ord ranked #5 on returns for the S&P in 2004.
"Timer Digest" has ranked Tim Ord as the #2 gold timer for 2004.
Market Analysis as of 1/26/05:
http://www.ord-oracle.com/Oracle22.pdf
OIH
Recently a fairly predictable trading pattern with good swings.
http://stockcharts.com/def/servlet/SC.web?c=oih,uu[w,a]dallyyay[dc][pb50!b200][vc60][iLi14,3!Lah12,2...
In addition to your mentioning of XLE, there is:
Utilities Select Sector SPDR XLU
Utilities HOLDRs UTH
Vanguard Utilities VIPERs VPU
iShares S&P Global Energy IXC
iShares DJ U.S. Energy IYE
A "must watch" on Friday's NBR:
Robert Morrow will be on Nightly Business Report on PBS tomorrow (Friday, 21). Unfortunately I will be out of town, picking up my latest and greatest computer (and having a few beers). He is a "must watch" in my opinion. He is a market technician and has a proprietary mathematical tool related to wave theory. I recall that in his early days he was a mechanical engineer and has numerous patents (perhaps related to vibration analysis).
I last saw Morrow on NBR in early in 2004, during which he predicted a low for the Naz of about 1,800 (which sort of pissed me off at the time because the market was much higher then) followed by a high of 2,400 just as the Naz re-enters a Bear phase. Here is a link I found regarding an NBR interview with him (article dated June 21, 2004):
http://www.thetechbriefing.com/issues/TTB_062104.pdf
Use the search tool to find his name once you open the PDF. Here is an excerpt from that article:
"Analysts K.C. Grainger and Bob Morrow, while continuing to be bullish about the tech recommendations they've made in past columns, including Advanced Micro Devices (AMD), Yahoo! (YHOO), and Qualcomm (QCOM), believe those stocks, among others, will suffer brutal declines in the not-too-distant future. They believe that, as in the past, the spectacularly performing tech sector cannot help but decline sharply in the expected S&P 500 bear market."
" 'We will have to change horses soon,' says Morrow. 'The tech horses should lose their power a bit later this year. You certainly cannot complain about the performances that the tech stocks have given us, but the end is in sight.' "
I also found this link on SI. It is an earlier transcript of an NBR interview on 03/12/04:
http://www.siliconinvestor.com/readmsg.aspx?msgid=19912003
During that interview, Morrow states:
"Well, I think the bull market is going to extend on into October this year [2004]. That will be the end of the bull market. The numbers I'm looking for on the Dow are 12,615; the NASDAQ Composite 2,378; and the S&P 1,345. That's the end of the bull market, in my view."
Is he right about the "end" of Da Bull and not so right on the numbers? He was very close on predicting the low of the Nasdaq.
Sorry, I also appreciate all the work you have done to present this data, but I am not impressed with the indicator. The times when the 10% line is crossed are mighty sparse. I would sooner crawl across the desert on hands and knees looking for a waterhole from which to drink than I would want to sit at my computer waiting for a PAT event.
Normally (and this is a major point, so it deserves its own paragraph), by the time it happens, no one would have the money on hand to take advantage of it anyway because they would have already been tempted, by whatever psychological means, to buy into the market beforehand. I want an indicator that tells me when to buy during those dozens of great opportunities between a PAT event.
Consider this: if there is a choice of only one great indicator, it is far better to have an indicator that tells you when to sell than it is to have an indicator that tells you when to buy.
Right now, I'm really glad that I am 70% cash in this market and the rest is half SHort and half LONG. Gotta love being SHLONG! My X was right after all ... window shopping is great!
Regarding your "Dow Theory Bull Market confirmation" question: the Bull market we are currently in was confirmed in March of 2003. By definition, new primary market trends are confirmed at the beginning of their trends, not somewhere in the middle or at the ends.
I had an ongoing argument with some guys several months ago who kept insisting that the current Bull Market started in October of 2002, because that was the market low. One could say that was the beginning of the Bull Market, but only after the fact and not according to Dow Theory. Dow was smart enough to know that you will never know the low of a market until it has been confirmed -- that indeed, without confirmation, the next low may be even lower. If the next low is lower than the prior low, then all that has happened is a secondary trend change, but the primary trend is still that of a Bear market. One valid criticism of Dow Theory is that it is always "too late".
Confirmation of a new Bull market does not have to be in the form of a double bottom. A single bottom would do if there were also extraordinary high volume/price action. For most people I know who understand Dow Theory, that didn't happen in October. Another pattern which could be used to call a bottom is an inverse head-and-shoulders.
Dow Theory is used only for identifying primary trend reversals -- Bear and Bull markets -- and not for secondary trends, such as corrections. What we had last summer was a correction in the S&P and Dow. Keep in mind that during C.H. Dow's days there were only the Industrials and Transportation indexes. There was not a Utilities index, and there wasn’t an S&P and Nasdaq, either.
The problem with the definitions of primary and secondary trends is that the Nasdaq is a large market in itself and it has a high beta compared to the S&P. Because of the high beta of the Naz, last summer's correction in the other markets was a virtual Bear for the Naz.
I would like to introduce a new thought for rumination. What appears to be forming (or perhaps already two-thirds formed) after the quick advance seen in the market after last August's low is a modified (or sloppy) head-and-shoulders pattern. Modified or sloppy because of year-end trading effects, in particular the typical low volume during the long holiday season. Could this portend a correction this summer in the Dow and S&P, and a summer Bear for the Naz? I think so, but not because of this pattern.
It definitely is easier to get a 10dma for an individual stock than for a huge group of stocks. One must be careful, though, to not apply the same trading strategy to an individual stock that was designed for a large group of stocks. Also, I question the validity of any trading system -- particularly one that only uses ten-days of data -- during the heat and passion and volatility of earnings season, especially an earnings season which is a precursor to the dreaded and historical slow time for technology stocks.
The only real positive I saw in Friday's action is that it was at the beginning of a three-day weekend. That could be explained by traders (window shoppers) finally loosing patience, ex vi termini, and pushing the buy button. The negatives are that the volume was below average (while the sell-off volume has been above average), and virtually all the recovery on Friday was accomplished in about as fast a time as the sell-off on Thursday -- about one hour at the end of Thursday, and one hour at the beginning of Friday. The rest of the time on both days was pretty flat. Doesn't show a lot of conviction.
Dow Theory isn't limited to requiring double tops and bottoms. More generally, it needs confirmation of a change in trend; otherwise the price/volume action is simply the continuation of the existing trend. The type of confirmation depends on the type of top or bottom that is being formed.
I don't quite understand what you are trying to show with your charts. Mathematically, one cannot use a chart showing the percentage of stocks above their 50dma to determine the percentage of stocks above their 10dma. One has to calculate it from raw price movement data for the universe of stocks. Unfortunately, there is no public 10dma (that I know about).
The Idiot's Guide to Gold Prices:
Put a mirror pointing backwards on the recent peak in gold prices then look in the mirror back to the middle of last summer. Given that the Tech market will suck by then, gold will be back at $380, and we can jump all over KGC at $5 like ducks on June bugs. We're half way there already.
Why not?
Based on what I see in that chart, your "EXTREMELY oversold" is a bit overstated. I would go for simply oversold. For day traders, your definition would be adequate I suppose. Extremely oversold would be when the oscillator has made somewhat closely spaced trips into that oversold area. As I am not going long anything right now, any quick spikes out of your oversold area will be opportunities for me to place additional shorts on the tech sector.
If NEM then why not KGC?
Stand to make more -- gold stock -- same price pattern -- higher beta.
Short-term bottom then back up to 32 then turn down again toward 30? Repeat pattern till end of January, then retest last summer's lows? Reads like a recipe for biscuits. There is a famous contributor on MarketWatch (who used to contribute to the SFChronicle) who thinks the semis have to correct a lot lower than that. PEG will suffer when there's no EG. Y/Y comparisons wont be good this year. Better to be safe than have EG on your face?