Page 1 of 1
Trend Following with EMA + CANSLIM
Posted: Sun Jul 18, 2004 12:44 pm
I have been doing extensive "back testing" on this year's stocks,
plus ALL the strong performers in O'Neils's books, and I have found that:
Instead of looking for the ever perfect cup with handle, flat bases,
etc, it is possibly better to look for:
(stocks that have high IBD ratings)
and use 50-day Expontential MA and 75-day EMA to determine entry. When 50 day crosses 75, BUY. When 50 crosses back down thru 75, SELL.
I have played with multiple EMA's and 50/75 works real well,
ASSUMING you are entering the stocks for the first time just after
the market turns to "Bull" (as called by IBD and also via chart
reading......last time was March 2003).
Investor Jones sees that the market just went to "Bull"/rally. It
is March, 2003, and Jones finally is ready to invest after having
studied chart reading and O'Neils books.
Jones starts screening for 50/75 crosses.
Every day he gets a list of them, then selects only those that meet IBD criteria. Only the best ones that meet IBD criteria (preferable top 1-10 Industry Group), he buys.
He sells when 50/75 cross occurs to the downside. THIS MAY TAKE 1-2 YEARS, and will likely coincide with the next Bear. If he wants, he can then short-sell in the new Bear, using same 50/75 cross.
REMEMBER, we are entering 50/75 crosses in a new Bull, not 1 year into it, when the "youth" of the Bull is not there. We are doing this in a new, fresh, healthy Bull.
No more "pivot points", cup with handle, volume dry up, breakout,
bases on top of base, false breakout, etc ad nauseum.
Jones is not a rocket scientist and needs something simpler.
I ask the audience to back test this with me, just in case I am
loosing my mind.
Posted: Sun Jul 18, 2004 11:19 pm
"I have been doing extensive "back testing" on this year's stocks,
plus ALL the strong performers in O'Neils's books, and I have found that..."
You are back testing on this years stocks? not sure what that means. However if u pick stocks that have performed well, such as o'niels lists, then backtest on those same stocks, the results may be misleading due to severe hindsight bias. You would need to get historical lists to test with, ie the actuall lists you would have had at that time u made the trades, rather than the current list of hot stocks. U may find that the stocks u see with hindsight are nowhere to be found on the old historical lists a good portion of the time.
"and use 50-day Expontential MA and 75-day EMA to determine entry. When 50 day crosses 75, BUY. When 50 crosses back down thru 75, SELL."
If u are picking high relative strenght stocks like the ones on oniels lists, they will often be in the middle of large runs, so if u are basing your entry only on the crossing of two long term averages, u may find u are getting in after a long, potentially very long, trend, if ever, on all too many occasions.
"I have played with multiple EMA's and 50/75 works real well,
ASSUMING you are entering the stocks for the first time just after
the market turns to "Bull"
yes once again "assuming". WIthout assuming, how would you have known to get in at that time. And, over the prior years of the bear/bull market, how many false starts existed using htat same method? I have in the past followed the ibd commentary, and i will say it is not much different from the other stuff out there: In hinsight they can always cherry pick their good calls, and leave things vague enough so that the failures are ambiguous and hard to determine.
"Investor Jones sees that the market just went to "Bull"/rally."
If only it was possible to have the wisdom of hindsight in the present. What signal would jones use to determine "bull rally?
These are things to consider. I like the basic idea of scanning for stocks and applying a trading model, in fact i use this approach. However such ideas can be more difficult to test than they first appear, or least i have found that t obe the case.
I contacted the w. O'niel co. in effort to get historical data of their indicators so that "hot lists" could be recreated for such testing. I was told that they do not offer such a product.
Posted: Mon Jul 19, 2004 12:17 am
Using the IBD rankings as a fundamental (rather than a technical/chart-based) screen of the entire population of stocks is very effective in narrowing down the field of potentially tradeable stocks.
However, I have problems with the O'Neil/IBD approach in that they do not provide precise rules for the exits. If I remember correctly they advocate a 7% stop loss and the exits are substantially discretionary.
This may be as much as most IBD readers want or can use but I think anyone reading these forums is capable of improving on a simple 7% stop loss.
Myself, I screen 1000 stocks (mainly small & midcap) based on fundamentals eg P:E ratios, Earnings growth, free cashflow etc and then trade this basket of stocks using technical analysis, nothing fancy just a MA crossover on weekly bars.
I only trade stocks on the Long side unlike futures which I trade Long & Short.
I only use margin in my equity trading when the underlying indexes are trending up strongly and only then for trades that originate between November & April.
Posted: Mon Jul 19, 2004 10:35 pm
I personally like moving average cross-overs. However, I'm dealing in a longer timeframe than you.
I've found that my results tend to be real good when I have the crossover as a technical signal. Use ADX (with DI+ and DI- confirming) and then look for price action to follow.
What that means is I look for a crossover, Check to see that ADX is rising and DI+ and DI- are in the right positions and price is above (or below for short) both MAs, then only buy if on the next day the stock trades higher than it's high today, or short if it trades lower than it did yesterday. Limit order.
My exit is a 2 ATR stop loss, or a cross of a MA that is 1/2 the paramaters of the entry.
You'll find on average that you get about 35% winners with about 4:1 in terms of average winner vs. average loser.
If I had the money to take all the signals I'd have made about 88 entries since the first of June and stopped out of 18 of them. Most of the signals lately have been shorts.
This method I found in Computer Analysis of the Futures Market by Lebeau & Lucus (gotta love trips to Portland to walk the shelves at Powells) which is a good read in terms of general system building and use of indicators.
Results From a Simple EMA Crossover System
Posted: Sun May 27, 2007 12:19 am
I noticed that most of the stocks the simulator picks are in the IBD 100 or end up in the IBD 100. The IBD 100 are the best stocks according William O'Niel's picking criteria.
Posted: Tue May 29, 2007 1:53 pm
What seems UNclear is what you meant by "ASSUMING you are entering the stocks for the first time just after the market turns to 'Bull' (as called by IBD and also via chart reading......last time was March 2003)."
Does that mean I can't enter on a 50/75 cross if it's the second such cross in that stock since 03/2003? If every stock on the IBD list has previously had an entry and an exit signal, I would be left with no entries?
Is there an objective method of denoting what a "bull" is?
Perhaps - not wanting to put words in your mouth, so this is my idea - use the 50/75 cross of the index as the start/end of each investing cycle and trade the first 50/75 cross for all stocks that meet basic IBD criteria (sector strength, relative strength, earnings growth).
[Edit - that 50/75 SPX cross would make August 2006 the last bull start, BTW]
Posted: Tue May 29, 2007 3:46 pm
It would be nice if IBD offered their data. However, here's something worth noting in my experience with system trading stocks:
I use a long/short filter based on timing the S&P 500. If the indicator says to buy the S&P 500, I'll use a suite of trading systems that only trade long. When it says short, I use a suite of systems that trade short.
I've looked into combining fundamental data with long/short trading systems, but nothing came remotely close to the top down trading approach. I believe it's because stocks are so closely correlated to each other (obvious but very true in my testing and actual trading).
The trading systems just seem to boost performance of my S&P 500 timing (~70%+ CAGR vs. ~40%...MAR ~2 ). My conclusion is that the real edge in my trading method is really the S&P 500 mechanical timing. The other stuff is just icing on the cake really.
I do the usual for stock trading:
Buy breakouts, sell breakdowns, buy prior to breakouts, sell after n days, etc. Not just a trend follower. There's lots of ways to trade...
What's funny is that my testing shows I would do better just trading LEAP options on SPY or OEX rather than individual stocks. Makes me rethink how I trade quite a bit...
Posted: Sat Aug 18, 2007 2:24 pm
One of my best shorts was Doral Financial (DRL), the largest bank in Puerto Rico. It had a high Investors Business Daily ratiing in late 2004. Stock was in the $40's.
As I write it's priced at 89 cents.
Ironic that William O'Neil made a public remark he'd "never" made much money shorting. Yet, he wrote a book about the topic. He also owns a NYSE specialist firm. I can assure you they short each day, probably several times, though in the majority of cases, for brief periods.