Difference Between Screens and Simulations: How to Reconcile????

Research
1

Fellow Investors,

I have been working with screens for years. I use the AAII Stock Investor Pro, I pay $750 a month for the Reuters “Knowledge” product, Valueline, etc., and I’ve tried many others. Mark Gerstien (author of several books on screening) is my e-mail buddy.

But I recently discovered this site and I am very intrigued by the ability to back test theories and minor changes in those theories.

So I started with what I know: screens. I made a screen that is similar to what I have had for several years. It’s a combination of growth and value parameters. The back-testing capability here is incredible. Change or remove factors and see immediately what the result is over time…

But here’s my problem: I have tried to convert a SCREEN into a SIMULATION. I came down to a screen that had only 5 rules and worked best with the Momentum Value Ranking at 95+. It produced a chart with no time below the S&P since 4/2001 and 836% Total Return. Annual return averaged 209% per year with turnover averaging only 24%. Low turnover is very important to me.

When I tried to transfer those same parameters into the SIMULATION, it was a distaster with only a 30% annual return and a 686% annual turnover.

What gives! Can someone please help me?

ginger

2

A simulation is more realistic than a screen because it takes into account things like slippage, trading costs, and number of positions. These effect are magnified with lower capital. For example a 20 position portfolio will be completely different with 10K starting amount vs 100K. In the former a $10 commission represents about 2% of the trade, in the latter 0.2%. THe effect of this can be huge. I ran GARP top ranked with 100K then 1K leaving everything else equal. The first returned 590% (62%/y), the second 72% (14%/y).

Another source of difference is that a screen is always fully invested as long as one stock passes the screen. It buys all the stocks that pass equally weighted. A portfolio might leave cash on the sidelines if not enough stocks pass the buy filters.

3

One more thing, when you say “with turnover averaging only 24%”, where do you get this from? THe turnover column in the screen is the turnover from the previous period. In your case it’s weekly, so a 24% weekly turnover is very high, much higher than a 686% annualized.

4

Marco,

Gotcha. I definately see how a sim would be more advantageous in a real-time situation, but I cannot seem to go from a 10-rule screen that is generating 80+% per year to the simulation. They are completely different animals and my best attempt to duplicate the screen in a sim results in 23% annual returns.

The most difficult part is when I try to duplicate many of the Screen Rules as “Buy Rules,” it is unworkable. I can make a screen with 15 rules, but the sim only allows 6 buy rules.

I would like to try a sim without a ranking system and just use my prop screen with more than the allowed 6 buy rules - plus the advantage of stop-loss/max loss sell rules. Is this possible?

Also, you are right about the turnover number I was looking at. It’s the weekly average for 3 years not the annual turnover.

Chris

P.S - You can delete the lower one of these two duplicate posts. I accidentally hit send twice. ;0)

5

You can have more than 6 buy rules in a sim. Just combine several buy rules in the same Buy Rule field using “AND” or “&”.
For example, you could but this into one field: “EPS%ChgTTM > EPS5YCGr% AND OpMgn%Q > OpMgn%TTMInd”.
You can do this for sell rules also, but it changes the logic. If you put 2 statements in the same sell rule field and connect them with an AND, it would not sell unless BOTH were true. So for sell rules, you would want to connect them with an OR (unless you actually wanted to sell only if both were true - then use the AND).

6

Dan,

Thanks for the suggestion. I was just taking things at face value and hadn’t considered that you can combine buy rules. That is not a feature that is commonly available in most screens and is a definite advantage in P123. Chalk it up to an experienced “screen-user” that wasn’t aware of that feature available in this site.

However, it is a VERY IMPORTANT FACT that the BUY rules are all “AND” rules and the SELL rules are all “OR” rules. I have looked through the guides and this very, very important message did not get through. I’ll read them again, perhaps I missed it because I would rather dive-in than read, a method I’m sure many can relate to…

I am still in the early stages of the learning curve with P123 - but I believe Marco has made an excellent product for both the very sophisticated investor (or professional) and the amateur. This site is definitely not for amateurs if they are going to run complex sims and try to create advanced screens (in my opinion).

But the value is excellent for those amateurs that simply follow the best screens created by the public and are shared in the weekly e-mail updates. The people that have been studying the market for years and share portfolios/sims/screens make it easy for those that are new to simply follow the “best-of-the-best” ideas. The facts and profits will out over time…

It was a stroke of genius to come up with a site that allows the public to experiment and share their creativity in a venture as complicated as “beating the market” - and truly something that would never have taken place but for the Internet. Kudos to Marco for the idea! It is world-wide democracy at work and only in this environment would everyone be willing to share their knowledge, experience and creativity with everyone else at no cost.

I look forward to learning much more from those with a different creative approach than me and also contributing whatever I can that is worthwhile to the publicly-shared creativity. This site is new, so it’s only going to get better as more minds are put to work and contribute.

I’ll certainly be giving some editorial page-space in my website and print publication to this excellent idea and Portfolio123. Thanks Marco (and all site contributors)!

Christopher Michaels
IntelligentValue.com

7

Ginger,

The real power of this site is the Ranking Systems. Unlike Screens, were each stock MUST meet the criteria of a Screen, the Ranking Systems “Rank” each stock with respect to each criterion, and then rank them with respect to all the criteria.

To use this tool effectively you must re-think your approach to Screening. Instead of using your screen criteria in the Buy and Sell Rules, develop a Ranking System based on your favorite screen rules. This will allow a stock that barely fails one of your screen criteria, but has the best values with respect to the rest of your criteria to still be ranked, and it will have a very high ranking. It could be the number 1 stock in your screen except for that one factor that fails your screen. With your screen, you will never know it existed!

Be sure to use the Ranking System Performance Page to test your System and try different Factors to optimize your System. I add factors one at a time and check the performance to see if it helped or hurt. If it hurts the performance I remove it even if it was one of my favorite screen criteria. You will be surprised how poorly some of your screen criteria perform.

After you develop your Ranking System when you put together your Sims use only the few screen rules that you want to be absolute in the Buy and Sell Rules. With a good Ranking System, you should only end up with 3 to 6 Buy or Sell Rules.

Remember: “The power is in the Ranking Systems”

Denny:sunglasses:

8

Denny,

Thank you (and Marco and Dan) for your advice to someone that acknowledges the incredible apparent power of this site, but is flustered that many years of learning must now be turned around in a different direction. The number of hours I spent developing and back-testing screens must be re-formatted into a new, apparently more powerful system.

I completely understand your suggestion that a screen may barely miss an excellent stock, while this system may not.

Such is life. I feel like an early astronomer at the time that Galileo said “The universe does not revolve around the Earth.” Hmmm… time to re-think everything (perhaps).

I have been very successful in the last 7 years with my screens that I used as a “watch list” and which resulted in average annual returns if 160%+. I see this site as an excellent tool to develop even more powerful buy/sell decisions.

I publish a newsletter called “IntelligentValue.” You can check out my performance at www.intelligentvalue.com. If you look at the performance graphs, you can see that from Sept. 2004, when I first started publishing, to 12/31/04, things were hunky-dorry and going according to plan (and in line with the last 6.7 years of tracking results). But as many of the top Sims (including yours, Denny) are showing, this year and especially March 05 has been very, very bad for value-oriented systems.

I have made “executive decisions” to pick stocks that would over-ride the usual parameters of my screens to buy energy stocks in order to continue beating the market in the last three months. Energy is up 15% this year while it is difficult to find anything else that is working.

As a new publisher, I am under constant pressure to make money for subscribers. A new subscriber doesn’t care about what I’ve done before - he just cares that he lost 10% since he subscribed two weeks ago (as an example - hasn’t happened yet).

Therefore it’s always a challenge to beat the market. One interesting aspect of this site is Marco’s inclusion of an e-mail summary each week that tells of the sims that beat the market in the last 1-week and 4-weeks. I would not change my strategies completely to try to match the “latest and greatest,” (plus, the strategy may not work next week), but the advantage to learn from what works in different economic environments is excellent!

I would interested in paying a “consultant” to help me develop some sims here based on my successful screens (with the approval of Marco). Denny? Mitko? Marco?

Let me know if anyone is interested…

Chris Michaels

P.S. “Ginger” is my username and comes from my first pet, a beautiful black and white cocker spaniel that my parents got for me when I was two years old.

:slight_smile:

9

Chris and Ginger,

If you were already averaging 160% per year, I wouldn’t mess with things too much. It might lead you into a state of confusion and indecision, and we know what indecision does to a trader.

Your existing screen does (might) sound like it does the job, but the screen is only a small part of a complete system. The exits are probably more important than the entries. Research has shown that a person can make money by picking stocks at random, as long as he/she has a sound exit strategy. Have you ever tried using your system with randomly-picked stocks? That might lead to some interesting results. That way you would know how much of your success can be attributed to you screen, and how much to the rest of your trading system.

I am also new to this site, and am very intrigued. Many sims show an annual return of over 200%. But this may be very low, since they are only rebalanced weekly, and many stop losses can slip greatly in a few days. If stop losses are rigidly followed, these returns may be much higher. I would like to hear actual case stories from people that have been here for awhile. Marco - maybe that can be a new forum subject?

In any case, congrats! It sounds like you are doing very well.

Brian (traderblues)