| Index | Recent Threads | Who's Online | User List | Search |
|
|
 |
Portfolio123 » List all forums » Forum: General Comments » Thread: PNCL - What to do now? |
|
Total posts in this thread: 31
|
[Request a Feature]
[Post new Thread] [Add To My Favorites] [Watch this Thread] |
| Author |
|
|
nrhp99
Advanced Member
|
I am fortunate to own PNCL, picked me up from a down week and gave me smiles and happy feet with the over 50% gain on Friday. Can one trust the P123 ranking system in this unique situation as my port is saying keep PNCL after rebalance, but I am thinking that the latest fundamental data may not have been made available to P123 with such a fantastic gain on Friday. What are your thoughts? |
||
|
|
gviersen
Advanced Member
|
I checked - Pinnacle didn't report fundamentals or release earnings - only its restructured 10-year deal with Northwest, on Friday. That would contain no fundamental data that would be updated either on Friday or this week.. Therefore - don't expect your rank to change, at least due to fundamentals. Technicals or valuation factors may change with the price increase, but that should have been included in today's rebalancing, since price data was downloaded overnight. Pinnacle's fundamental data probably won't change for awhile. It is a calendar-year company that will release Q4 2006 earnings approximately the first week of February (based on historical dates). Even then, a vast majority of fundamental data may not get updated (this is when "N/A" v "Use Previous Quarter" is important on P123). Companies control the data they release at an earnings releases - it may only be earnings, it may be earnings and revenue, or it may be a more extensive set of data. On P123, the data that companies choose not to report then becomes #NA data (until the 10-Q or 10-K is filed, see below.) The full fundamental data update does not occur until the Form 10 filing through EDGAR with the SEC. Unfortunately, the annual filing at end of fiscal year (ie, 10-K vs a 10-Q) sometimes take months to prepare. Many calendar-year companies like Pinnacle may not file 10-K's until May, a week or a few days before releasing or submitting their 10-Q for the NEXT fiscal year (fundamental ranks really do a mambo in late April and throughout May). What does that mean? It means, depending on the fundamental data you use in your ranks, it may be four to five months before you get a fundamental data update for Pinnacle. If you are using data that Pinnacle does not release in its traditional earnings' releases, it will be April/May 2007 before it gets update from the last 10-Q filing in mid-November 2006. Oh-the other fun news is that small-caps like Pinnacle, especially calendar-year small-caps, are further down the priority list at Reuters when it comes to updating fundamental data. Like any business, Reuters serves its biggest clients best and larger companies have more analysts at more prestigious I-banks who pay Reuters more money. When it is busy at earnings' season, particularly with calendar-year companies (~70% of cos. are calender-year), the data entry folks and analysts at Reuters are swamped. I've seen it take 3 weeks for some of my small-cap data to get updated after an earnings release. I will say, however, that because the Form-10 filings are highly automated through EDGAR, the full 10-K or 10-Q updates tend to come through faster. So...tough call. Don't know you or your risk-tolerance or system, my only advice is what I do when something like this happens (rarely 50% in one day - good call on that one - when I get a one-day spike or a stock up 50%). I would get on P123 this weekend and simulate your system (or backtest) with a sell rule like price/price(1) -1 > .30 (or .40 or .50) and mktcap > 250 (or 100 or 300, etc.) See if it helps, hurts or is neutral to your normal system. 50% price increases in one day are pretty rare for non-penny stocks so you may need a lower threshhold to get any cases. (The market cap criteria is to keep out penny stocks from causing aberrations in the results. ) If you are a true empiricist, you will go whichever way the data tells you. I am a semi-true empiricist, so here is what I do when a stock gets to 50% gain. I sell down a third of it, back down to my target position size. Don't forget, with the 50% increase PNCL will have much more weight in your portfolio. If it craters after the big spike (which it will if the spike was caused by a short squeeze rather than true buyers) at least by taking a third out ASAP, you will have taken some money off the table and lowered PNCL's impact on your portfolio's returns. The biggest problem may be that the drop may happen pre-market or at-the-open on Tuesday. Cannot test this strategy in current version of P123 (cannot sell partial lots), so don't know if it is long-term maximizing of profits. I just know I sleep better at night knowing I've already taken the profit off the table and that the remainder can halve before I would actually have a loss . My stop-loss rules would take care of that in most situations. Sorry for the long response - let me know if you have questions. ---------------------------------------- [Edit 2 times, last edit by gviersen at Dec 23, 2006 11:29:48 AM] |
||
|
|
gviersen
Advanced Member
|
MARCO: This reminds me to ask if you have thought about implementing the fix you mentioned early in response to one of my posts. The fix about being able to set the parameter "Use N/As" or "Use Previous Quarter" separately for Buy Rules vs Sell Rules. As I mentioned, due to momentum, I would error on the side of "Use Previous Quarter" for sells, but would only want fresh data ("Use N/As") for buys. Crazy time is about to start for 2006 calendar-year end fundamental data and would be wonderful to have in place by mid-January when Alcoa kicks off the earnings' season. (Hard to believe a company that huge can completely close its accounting books in < 5 days at the end of a quarter. Blows my mind.) |
||
|
|
p123robert
Advanced Member CANADA Joined: Nov 7, 2005 Posts: 372 Status: Offline |
My thoughts are pretty much the same. YES, when you have a gain like that, it would feel a LOT safer to sell! I wonder if Reuters has had time to update EVERYTHING #1: This is just a small cap stock (and Reuters sometimes updates small cap stock only the week after!), and #2: This change occured on a Friday, the last day of the week! #3: This change occured before a major holiday! UFPT is an example that comes to mind. A few months ago UFPT was flying high, and UFPT was my best stock. Then UFPT was starting to go downhill, and gradually lost it all. Then UFPT became the biggest loser. When I sold UFPT, it was the biggest loser! However... NO, we shouldn't sell, because computers don't make mistakes, but we, humans, do! I hope this helps. Robert |
|||
|
|
tyldak
Advanced Member
|
I decided to sell after that pop. I have no idea if that was the right thing to do, but so often it has been the case that I get a big, fast gain on some small cap, only to watch it drop almost all the way back to where I bought it, that I decided to grab the money this time. I had a 65% gain, and it's kind of hard to argue against taking that off the table. But, I may be kicking myself for that a month from now. We'll see. |
||
|
|
scottkissinger
Advanced Member
|
I also held PNCL and decided to sell 15 minutes after the open on Friday as I saw it start to come back from the highs. I am kicking myself somewhat for "going against the model." Even today when I went to rebalance the port that held PNCL, it picked PNCL again (no LTSellDays in that Port). I just went onto the next recommendation. After a power spike like PNCL had on Friday, you sometimes see pullbacks. And, sometimes the price continues to go up. Who knows if it turns out to be the "right" decision or not....time will tell. In another week I may be buying it back if the model says so, but I decided to lock in some gains now. But, this is certainly a nice problem to have....  |
||
|
|
IntelligentValue
Member  UNITED STATES Joined: Oct 17, 2006 Posts: 4 Status: Offline |
SCALING OUT YOUR POSITIONS What you are talking about (selling portions of your winners) is called "Scaling Out". gviersen mentions in his first post that P123 cannot sell partial lots in order to scale out of a winner, but it can. You should use the "Eval" function like this: Eval(GainPct > 30, 0.25 , 0) In the Rule above, the Eval function looks at the first condition (GainPct > 30) and if it is true, it invokes the first expression (sell 25% of the position), otherwise, it leaves the stock alone. I'm not sure this is a successful approach though. Selling portions of your winners seems like a logical thing to do, but in practice it has proven to be a money loser. There is an excellent article on this approach ("scaling out") at http://www.rightline.net/education/battle-of-exit-strategies.html. Instead, why not sell your winners when they have reached a certain profit level, then let P123 pick the next winner for you based on high ranking in your Port? LOCKING IN YOUR PROFITS In my online newsletter at IntelligentValue.com, I have recently implemented an "tightening TSL" policy for situations like the one with Pinnacle (PNCL). When a stock is first bought, the TSL (with the broker) is set at 18%, which is a little more than the simulation's backtested TSL of 15%. Then if a stock has gained 10%, the TSL is tightened to 10%. If the stock gains 20%, we tighten the TSL to 5%. If the stock gains 30%, we tighten the TSL again to 3%. The reason I went to this policy is that I was very frustrated when watching a stock gain 20% or 35%, etc., then watching it go all the way back down - sometimes to the set-in-stone 18% TSL that I had on all stocks. So I had the programmers at FinancialContent.com create an adjustable TSL system for my portfolios so they could be more like broker capabilities in the real world. So far, this system has worked extremely well in the new "Power-Value" mid/large-cap portfolio that I started on September 25. This adjustable TSL system has locked in excellent profits on many stocks, such as 22%, 33%, 25%, 18%, 46%, etc. You can see the all of the Closed Positions of the portfolio at http://www.intelligentvalue.com/pv_closed_positions.htm. The portfolio has gained about 25% in 3 months or about 150% annualized. You can check out the Power-Value portfolio results at http://www.intelligentvalue.com/results.htm#PV. Remember that this is a mid/large-cap portfolio holding stocks with higher trading volumes and prices over $10, so it should make much lower returns than a small-cap portfolio - but it's not. Unfortunately, this policy is not something that can be backtested in P123. There is no way to set a TSL that emulates the daily high/low price assessment of stocks like a broker would, while still keeping the weekly or monthly rebalance of our systems in place. Running a daily rebalance will not produce the same results as this is producing in the real world because it will sell daily based on other sell rules - such as rank - and not just the TSL rule. In my dream world, we would do away with the global rebalance command and the buy and sell rules would have a drop-down so that you could test invoking each one at different time periods. This would be most applicable to TSL and technical analysis rules (i.e., SMA(50)>SMA(200)... or Vol10DAvg>Vol3MAvg) or any rule that is price-related. I would love to see this function implemented but I'm not sure how many votes it would get. Is anyone else interested in being able to mix daily and weekly rebalances for different rules? Christopher Michaels IntelligentValue.com ---------------------------------------- Christopher Michaels, Publisher IntelligentValue.com "Almost an Unfair Advantage" |
||
|
|
crakes
Advanced Member
|
I agree with those who say this is a nice problem to have! I ran several simulations on what I am calling "Jumpers." These are stocks with big one day gains. I ran simulations on my trading system that got me into PNCL and found that there is no negative affect on selling Jumpers that jumped 30% or more in a day. In fact, it is very slightly positive to do so. It is so slight because in the trading system's 5 year history, there were only 3 jumpers of 30% or more. Meanwhile, Jumpers of 20% or less, which are relatively frequent, have actually been good to hold in my system. Then I wanted a more general sense of how Jumpers do. I ran a test with no ranking system (someing I built called Zero Rank helps keep ranking a non factor). Once I eliminated penny stocks, I found that Jumpers tend to do very poorly as a rule when you hang on to them. The last series of simulations I ran using a Value ranking system with no price momentum factors. Jumpers tended to hold their own, albeit with slightly negative bias (no commissions/slippage used) when sold 5, 10, 20 trading days later, but showed a more noticeable negative dive when held more than 40 trading days. The chart did however have periods when the Jumpers seemed to keep going up by a significant percentage. But alas, not enough to offset the more generalized downward bias. In summary, my trading system does not rely on Jumpers to make money. And in general, Jumpers don't do much after their jump. But then again, some periods showed that sometimes jumpers kept bounding up... Finally, there is the well worn statement that system traders should always stick with their system. Glad I could add clarity to the discussion ;-) Thanks, Carl |
||
|
|
gviersen
Advanced Member
|
Unfortunately, that rule does not work because it will continue to sell down the position every week, as long at the stock stays over 30% gain. No way to do a one-time sell of a partial position. I do like the idea of using different periodicities for each buy/sell rule. |
|||
|
|
tyldak
Advanced Member
|
I haven't tried it, but will this do the trick?: Eval( close(0)/close(5) > 1.3, 0.25 , 0) |
||
|
|
[Show Thread Printable Version] [Post new Thread] |
PNCL - What to do now?
