Trading Stocks in a Broker account for a P123 Model

Research
1

All:

What is the best way to buy/sell stocks to rebalance a portfolio? There have been some recent comments in other threads on techniques that members use. I am starting a new topic to focus the discussion which i think is important to the real $ performance of a model. Maybe there are others that would like to share their experience. I will start.

Stop Losses: Denny recently commented about the use of stop losses - he sets his trading account stops at 5% higher than the model. I started using a straight $ stop (say 86% of the buy price) then switched to a trailing % stop (14%) when I realized this is a better match to the model which was optimized to a 14% stop. My models are rebalanced weekly so this means the model stop losses are checked over the weekend. I found that I was being stopped during the week occasionally when the model did not show a stop sale on rebalance at weeks end. I now use a higher trailing stop in my brokerage account (20%) to try to avoid this problem. This particular problem cost me about 10K when KNDL took a one day dive on an changed analyst recommendation - I got stopped out, the stock recovered in one day and I had to repurchase at the higher amount. (I chose to repurchase since the stock still had the higest ranking compared to others recommended by the model)

Buys: I use limit buys so I do not buy all the available volume at any price. I found a straight buy would result in higher prices paid. I have been looking at setting a limit buy slightly above the close so I could buy when the market opens. It seems to me that the buys seem to increase in price from the previous close. I look at the bid/ask and try to buy at the last price but will set a limit above the last, and even the ask, to buy a rising stock.

Sells: I also use limit sells for the reasons above. I typically sell and then buy to rebalance. You get the $ from the sell and then buy the amount of the new stock at the current price. I do this mostly to try to match the brokerage account to the P123 model. I did not use margin. I am thinking that a better system is to buy what the model recommends and sell the models recommendation at the same time since it seems that the stocks I sell are mostly declining and the buys mostly increasing. i added margin to my accounts that allows me to do this. Any ideas? Does anyone have any experience placing buys and sells before the market opens on Monday?

Trading Costs: Since I am running models with 2 and 3 stocks, I tend to buy $10,000 or more. At $7 a trade the trading cost is a small part of the cost of the trade if you use the .5% slippage. To me, efficiency is more important.

Hopefully, the above may help someone that is starting out. I am sure others with more experience can add techniques that have worked for them.

Glenn Fagerlin

2

Glenn,

This is what I am currently doing:

Stop losses: Currently I am doing the same thing you are. I am setting a trailing stop of -20%. However, I am starting to do more daily rebalancing and using an ATR trailing stop. For this I will set a trailing stop equal to 5 times the 10-day average true range. For example, if the average true range is 40 cents, I will set a trailing stop of $-2.00.

Buying: At the start of the day, I set a limit order equal to 1% over the previous day’s close. This will catch most stocks, and many times I will get a price much better than this. If the stock has not been bought by 10:00 central time (1.5 hours into the day) I will modify this order slightly and use a bracketed buy. I will still leave the target buy price where it was, but will add a buy stop. This buy stop is 5 cents higher than the high so far for the day. This will prevent stocks from getting completely away from me. If it still hasn’t been bought by the end of the day, I will just set a market order a few minutes before close.

Selling: Just a reverse of the buying procedure. For example, I start the day with a limit order 1% below the previous day’s closing price. After 10:00 I will change it to a bracketed sell.

I have been keeping track of the average slippage using this method, and it is running at about .75%, trading small caps.

Brian

3

Brian
Brain-dead me would like to know what is/how is a bracketed buy. The openness and generosity of all is continually to be applauded and is graciously received. Stephen (Ireland)

4

Stephen,

Some brokers offer advanced order options, more than your usual market, limit, or stop orders. Check with your broker to see if they have advanced orders.

A bracketed order(some brokers may use different terminology) is where you give two different prices at which to buy. The main price is your target price (some call it a trigger price), or the price at which you would like to buy the stock. The second price is the stop price. This means if the stock climbs above this price, you want to go ahead and buy the stock so it won’t get away from you.

For example, say a stock is currently trading at $8.50. You would like to be able to buy the stock at $8.25. This is your target price. If the price drops down to this level, your order will be triggered and sent to the market maker as a market order at that point… So this is similar to your limit order, but not quite. However, you think that if the stock goes above $8.75, you want to go ahead and buy it before it goes to $10.00, so you set your stop price at $8.75. If either of these prices are hit, and the order is executed, the other price will be cancelled automatically.

Anyway, that is the way my broker handles it. Others may handle it a little differently. My experience with these bracketed orders, though, is they don’t seem to get filled as quickly as regular limit orders. At least not with my broker. That’s why I start with a limit order.

Brian

5

Brian
Thank you for the clear and lucid explanation
Stephen

6

Hi Glen,

Stop Losses: Just recently, Denny made an excellent argument for using stop losses. Additionally, I get the feeling, most P123 members use stop losses.

That’s fine. Feel free to do what works for you!

If you want to benefit from my expereince., here’s what I do. I myself completely avoid all stop losses. Why?

A ) Because of bad experiences. In the last 18 years I’ve had bad experiences with stop losses. Experiences that were similar to the one you’ve described (i.e. the KNDL down spike that cost you $10K, when KNDL took a temporary dive, you got stopped out, and the stock recovered the same day, and you had to repurchase it at the higher price). The problem is, all you need is a few $10K losses due to down spikes, and, for the rest of your life, you’re going to be nervous about having live orders out there!

And B) because, in my experience, the probability that the stock recovers is high. AND the probabilty of a temporary spike is also high. Therefore, why use a stop loss, if the stock recovers just about all the time? And, why use a stop loss, if a big down spike is also likely? Remember, those big down spike will sell your stocks at a ridiculously low price, at 20-30% below market!

And C) because, stop losses give you a false feeling of security. One problem that I see is a non-transparency that is associated with big down spikes. Then, to make things worse, those big down spikes are removed! Ostensibly, to make those stock charts more readable! Yes, it’s very nice that the charts are cleaned up cosmetically. And, yes, it is very nice that clean charts with little or no volatility make investors less nervous! However, with electronic trading, big down spikes are just around the corner! You’re going to feel secure till the next big down spike confiscates $10K of your cash! And then, when you’ve lost $10K, your loss is going to be ignored everyone, including ALL market data reporting services!

And D) because the things that occur in connection with spikes is a little disturbing to me. I’ve already told you about the cosmetic cleaning of charts and daily data, which practice I find questionable, because it’s a lot like not telling the truth, because people do get stopped out at those extremely low prices, and real/actual trades do take place at those low-low prices! In addition to this non-reporting, I also find it disturbing and unfair that whenever there’s a big DOWN spike, your stock is SOLD at that LOW price! But, whenever there’s a big UP spike, and whenever you’ve got a limit order to sell your stock at a HIGH price, you stock will NOT be SOLD!

However, if you feel comfortable with A, B, C, and D, feel free to USE stop losses!

Do whatever works for you!! I’m only giving you my hard-earned experience of 18 years in the market.

I hope this helps.

Robert

7

Hi Robert,
thanks for sharing your experience with stops.
What you are saying are all legitimate concerns.
Unfortunately there is no “perfect” solution for the dilemma we are facing in using / not using Stops.

It has been my experience that the concerns you are referring to are often (not always) associated with poor liquidity or small cap stocks. The stock referred to in the example was KNDL, which is not particularly liquid (only 23,000 shares traded on Friday). I found many days of under 100 k volume for this stock. I tend to avoid these stocks.
I know this CAN happen to other, liquid stocks as well, but I think it is not often enough to completely abandon stops.

Example: I had a portfolio of 5 stocks on 9/11 with a 20% stop loss. 4 of those were stopped out at 20%. But later in the day, 3 of those were down 40% and closed near that level. So my stops have indeed protected me from a much deeper loss and this was by all means a “catastrophic event”. This is exactly why I am using stops: To be prepared for that rare but “catastrophic” event, which can wipe out a portfolio in no time and devastate your financial future.

I know, this solution is not perfect but NO protection is not a viable option for me either.

Cheers,

Wern

8

[quote]
thanks for sharing your experience with stops. What you are saying are all legitimate concerns. Unfortunately there is no “perfect” solution for the dilemma we are facing in using / not using Stops.

Example: I had a portfolio of 5 stocks on 9/11 with a 20% stop loss. 4 of those were stopped out at 20%. But later in the day, 3 of those were down 40% and closed near that level. So my stops have indeed protected me from a much deeper loss and this was by all means a “catastrophic event”. This is exactly why I am using stops: To be prepared for that rare but “catastrophic” event, which can wipe out a portfolio in no time and devastate your financial future. I know, this solution is not perfect but NO protection is not a viable option for me either.
[/quote]Hi, Wern,

You’re welcome.

Do whatever works for you! My problem is that, if I used “AvgDailyTotal > 200000”, as opposed to “AvgDailyTot > 100000”, my sim’s annualized return would drop significantly. Therefore I use “AvgDailyTot > 100000”, which means, from time to time, I have to trade low volume stocks. Because of low volume stocks, I have to worry about the next big down spike.

Naturally, I worry about “catastrophic” events as well. I was fully invested in the market in both 1989 and 2001 when the market dropped 30% and 20%, respectively. The shock and surprise we felt on both of those days is something we can never forget!

However, my decision-making is based on probabilities. What are the probabilities? “Catastrophic” events like these occur approx. every 4000 days, but big down spikes occur approx. every 4 days. Therefore I have to be 1000 times more protected against big down spikes. Therefore I don’t use live stop loss orders any more.

However, your situation may be different from mine. Therefore, don’t change anything, just because what I’ve said! Keep doing whatever works for you!

Robert

9

[quote]
Buying: At the start of the day, I set a limit order equal to 1% over the previous day’s close. This will catch most stocks, and many times I will get a price much better than this. If the stock has not been bought by 10:00 central time (1.5 hours into the day) I will modify this order slightly and use a bracketed buy. I will still leave the target buy price where it was, but will add a buy stop. This buy stop is 5 cents higher than the high so far for the day. This will prevent stocks from getting completely away from me. If it still hasn’t been bought by the end of the day, I will just set a market order a few minutes before close.
[/quote]Hi, Brian,

Thank you. Your info is valuable, because this technique of yours does work, because I’ve just done some back testing, and, on paper, this technique of yours did make a positive difference! On paper, I was able to buy at lower prices, on a few days when the market for my particular stock gapped up and therefore opened too high!

I believe this technique of yours does work with listed stocks, because with listed stocks, at the opening bell, there are one-price auctions.

With unlisted stocks, this technique of yours works, too, assuming your broker and NASDAQ do treat market and limit orders equally well. In my experience, they don’t.

Just my hard-earned experience in the stock market.

I hope this helps.

Robert

10

Robert,

You are correct in pointing out that most of the time that a stock spikes down that it recovers. However, I think you should consider the following:

Your 18 years experience with stop losses is based on pre-P123.

Using P123 Sims/Ports I am always able to improve the annual return and reduce the max drawdown by adding a PctFromHi stop loss. If the Sim is improved with a stop loss, doesn’t that mean that the probability of gains using a stop loss is better than the gains without a stop loss?

Although most of the stocks that the Sims sell due to a stop loss do recover, most of those take a significant period of time to recover. On average, the stocks that the Sim replaces the stopped out stock with have a faster gain than the stock that was stopped out.

Many stocks are stopped out due to a gradual loss and not a spike down and are replaced with a higher ranked stock. Higher ranked stocks perform better than lower ranked stocks.

If it works in Sims, why not use it in real trading?

That being said, you still have to do what you are comfortable with!

Denny :sunglasses:

11

This is absolutely the truth! The best system in the world will perform poorly if the trader doesn’t know how to, or doesn’t feel comfortable with, trading it.

12

Denny:

What range of stops have you found to be effective when using PctFromHi stop losses?

13

Hi,

My question is whether it is possible to simulate in P123 complete rebalancing of all the stocks in portfolio back to equal weight of their value for each holding periodically (say monthly) rather than selling out completely from lowest rank stocks and buying new high ranked stocks only.

I was wondering if others find this feature of being able to do periodic rebalance of entire portfolio to equal weight be good for risk management? Or is it better just to sell the low rank stocks completely and buy new high ranked ones?

The ability to rebalance to equal weights (to the level of fractional shares) without paying transaction fees is something that seems to be offered by FolioFn. I would really like, if possible, to set up 2 simulations to compare those 2 cases.

I am new to P123 and am really impressed. Learning a lot by reading the forums. My experience so far is mostly with Mutual Funds…am hoping I can use P123 to intelligently learn how to move more into stocks to target somewhat higher return while being careful not to lose risk management from diversification and asset allocation I have enjoyed with Mutual funds…I read through a lot of forum topics and couldn’t figure out answer to this equal weight rebalance question (sorry If I missed it and it was covered already).

Thank you for your thoughts on this.

Kurt

14

Kurt,

You can accomplish that by using the NoDays Sell Rule. For example, if you want to rebalance equally every 3 months add the Sell Rule; NoDays > 90. This will sell all stocks after 90 days and buy the highest ranked stocks that pass your Buy Rules.

If you don’t want to sell all of the highest ranked stocks that you currently own, on Step 1 of the Sim select Yes for; “Allow sold stocks to be re-bought at current rebalance”. This will cause the Sim to re-buy any stocks that you hold that are still highly ranked, and it will buy the number of shares that equally balance the Sim. In your real Port you just buy or sell only the number of shares necessary to equalize the Port. There will be a small error in the P123 performance calculations due to the excessive commission and slippage on the shares that you didn’t sell and re-buy in your real Port.

However, in the Sims that I have tried this, it reduces the annual return due to prematurely selling some of the shares of the highest gaining stocks with very little reduction in risk. If you can come up with an approach that reduces risk without hurting return very much we would all like to know what that approach is!

Denny :sunglasses:

15

Thank you Denny.

Your suggestion enabled me to get going to make an initial comparison of with and without periodic rebalance to equal weight.

I see your point about how the rebalance to equal weight seems to reduce the gains (not good) without reducing risk (not good again).

Sampling two P123 rankings (Balanced4, Momentum and Value), with several different community simulations (small to large cap), with and without NoDays > xx (where 7 < xx < 90), almost invariably the returns were lower, usually much lower, when I added NoDays to the existing Sell Rules. And the draw-downs didn’t seem any better. So if one is allowed to define a return/risk figure–of-merit of Return/Draw-down, that figure of merit got much worse, not better.

I then tried removing some of the Buy Rules (the thinking being that maybe in some way they fight the NoDays Sell Rule) but I found those ports still had much lower performance when the NoDays command was present.

I also increased the number of shares significantly up to as much as 100 positions in a few different simulations (the thinking being that perhaps one needs a lot of statistics to start to see an effect), and that didn’t seem to help either.

This negative result is rather interesting and it is counterintuitive to me. Also, I have heard/read that rebalancing in general is a good idea (financial advisors often recommend this for portfolios of mutual funds).

Clearly this was a much too small sample set and not chosen in a systematic way since I am just starting to learn P123.

But, if there is a good benefit to this equal weight rebalance, it certainly doesn’t want to show itself easily…. or perhaps it only shows itself over many years time, when averaged over portfolios containing many stocks/bonds, etc. and not when invoked in the detail of a particular stock portfolio with a somewhat small number of holding with a particular set of Rankings and Buy/Sell rules….I am not sure…I need to learn more.

Thank you again for your help.

Kurt

16

Kurt,

The rationale for rebalancing to equal weight is based on the logic that you want your Portfolio to be set up so that no one stock can have a devastating affect on your overall investment if that company fails. This is purely a risk reduction approach. This is why it is recommended to have a minimum of 15 to 20 stocks in your Portfolio. With 20 stocks, a loss of 50% in any one of them results in a 2.5% drop in your investment. If one of your stocks increases to the point that it becomes 10% of your investment and then it lost 50%, you would have a 5% lose to your investment. Thus the recommendation to rebalance regularly.

However, if the Sims show that this is not a very effective investment approach, then why do it?
The P123 method of rebalancing ( periodically selling due to Sell Rules, and re-buying a few stocks at a time) will, over time, keep a fairly equal distribution of stocks, and they will have been sold due to a good tested reason and not just a specific time period.

Denny :sunglasses:

17

the Interactive Brokers’ TWS software has a rebalance feature that works well for this type of investing.

I am using this. basically what I need to do is to put in the ticker and the % of the portfolio, then it generate the buy/sell orders automatically with number of shares and limit price. The default is to limit buy on the ask and limit sell on the bid… but I think the default can be changed to something like limit buy on the bid etc.

I am not using stops… with a 20-stock portfolio, stops become sorta irrelevent.

18

Hello Denny and Dozu,

Thank you for your helpful comments and suggestions.

All, as I get further along on exercising P123 simulations and ranking systems, may I ask your guidance regarding setting up the SIMS towards goal of best can do representing real-life trading Buy/Sell experience. I have, for instance, questions of (1) choosing Price for Transcations and (2) Slippage.

Question 1: I am not sure what to choose, or the reason to choose which of following:

a) Previous close
b) Next Average of Hi and Low
c) Next Open

I note that the P123 Model portfolios use Next Open. But I also saw some forum posts pointing to Next Average of Hi and Low as being possible settings to consider. Much of the discussion about this seemed to take place in 2004 and 2005 and early 2006 forum posts, but I got a sense that there were changes made in P123 in 2006, so I wasn’t sure what to choose. What would be the reason to choose one setting over the other and why, to most closely represent trading realities?

Question 2: What is reasonable slippage numbers to use in simulations. I had read a number of forums which quoted slippage factors between 0.5 and 0.76, In case of FolioFn slippage seemed not-well characterized, based upon previous posts I read.

As background, after reading through the forum posts, I am considering as straw-man, to use 2 brokers, FolioFn (No transaction fee for window trades) for mid to large caps and Interactive Brokers ($1/trade, trailing stops) for small caps. But before I decide for sure, I wanted to practice with additional SIMS and ranking systems to get idea of how they might work before committing real $$. I am starting with small investment amount. So Slippage and how one sets price will be very important, particularly if go with weekly reblance, towards picking SIM to follow.

Let me say an immense thank you fo everyone for all of your insightful posts. Your ongoing advice written in the past couple of years in the forums has been teaching me very much.

Kurt

19

the price you select betwee open/close/average of hi and lo probably doesn’t make much difference.

I usually just use the open as that is about the time I execute real trades… in reality, I try to get the fills before 10:30 when liquidity is still good.

Slippage, it all depends on your execution… if I see a stock has good liquidity and bid/ask is close, I hit the ask to buy

if the spread is too wide, I will try to bid for a few minutes. if no fill, I let P123 make a new recommendation.

In my simulations, I use 1% to be safe.

20

Hi Kurt,

You want to use average hi & low or next open for sims with high turnover and/or momentum factors. There is usually some price jump over Friday closing on Monday morning for new buy recommendations (and some price drop for sell recommendations).

The results in a sim may be a little different using Average Hi & Low or Next Open. Most of time, the results are worse using Average Hi & Low. I believe it is for the same reason explained above. I use New Open for my sims as I do my trades on early Monday morning.