-------------------------------------------------------------- I posted the following chart to illustrate a rectangle pattern to another poster on Zeev's thread. --------------------------------------------------------------
-------------------------------------------------------------- Zeev then posted the following message to me. --------------------------------------------------------------
Augie, a good example of medium term "box" trading, with one difference, your first stop loss is set under the bottom of the last box (around 12.5), you raise that stop to just below the bottom of the new box, only after that bottom has been tested at least once (like just under $14.5). Then you raise it again to just under the bottom of the next box (now $15.5 to $16.75 or so, thus SL at about $15.25 (which should be just above your buy at $15?), and if $16 holds over the next few days you raise the stop to just under $16 and so on... Mind you, even though we may be no in a cyclical bull phase, quite a number of whipsaws should be expected. After all, it is not a secular bull....#msg-575769
-------------------------------------------------------------- I then asked Zeev a followup question --------------------------------------------------------------
So, when you sometimes take more than one position in the same stock, i.e., the day-trader/swinger combo, or the swinger/position combo, you are holding the longer position from one box to the next, (as you described), and buying- selling-rebuying-etc the shorter position as it rebounds off of box tops and/or bottoms. Is that a fair characterization?
-------------------------------------------------------------- And Zeev replied --------------------------------------------------------------
Yup, for swing trades, not for day trading, there you have three types of boxes, linear, ascending and descending, I play all of them....
------------------------------------------------------------- ------------------------------------------------------------- Meanwhile, over on SI, Steve Lee wrote a post to exp about Zeev & Darvas Boxes, -------------------------------------------------------------
Darvas used these boxes for multi day trading and did not advocate using them for day trading. Zeev uses a modified system and is obviously doing well with it.
The trading Darvas did involved using individual brealout stocks on the way up only. I don't believe, from what I have studied, that the system will work for an index. In fact the system was derived to find and profit from stocks that are acting in a special way, so by definition does not apply to the market. It also uses the volatility within a range, which again by definition, an index does not have to the same extent that volatile single sotcks do. Look at the way Zeev uses the Q, a quintet of particularly volatile stocks, for an illustration of this. They are not just all bought at the same time for their beta with the Nasdaq (although sometimes it happens that way). Sometimes Zeev sells one Q member while buying another.
As one Q member goes up and another goes down, the Nasdaq index can be standing relatively still and thus not presenting trading opportunities. #reply-18195791
------------------------------------------------------------- To which Zeev replied -------------------------------------------------------------
Steve, the pure Darvas approach should be used only in bull markets, it is designed to try and keep you in a stock for the whole bull move. One way to know that the trend changes is if in a short period of 4 to 6 weeks, more than 20% of your stops (set under a Darvas Box) are taken out, then you want to tighten the rest of your stop, or even get out all together except special situations. (all IMTO). Using the Darvas methos with one or two stocks only, assures whipsaws, it got to be used with at least 15 different stocks simultaneously, IMTO.
-------------------------------------------------------------- Steve responded --------------------------------------------------------------
re 15+ stocks for Darvas method.
I can see the logic here Zeev. Lots of small losses and a couple of big wins that more than compensate.
With only a couple of stocks you will just get a couple of small losses. Unless you are lucky. But a system shouldn't depend on luck.
I haven't tested but I think the system could work well for shorting also as a way of finding "broken" stocks.
I also feel that now we are in the days of mass trading from people using the net, levels of support and resistance are not so clear cut as in the old pre bubble fractional price days. So if Darvas were investing in the same way he would be getting more whipsaws. I guess this is where your tweaks to the method give you the advantage.
Anyway, I appreciate the comments because it is a system I am interested in and your system is obviously successful so any nuggets of info are gratefully lapped up (forgive the bolded pup.. er... pun<g>).
Steve, one improvement to the Darvas box is using it on a 5 minute chart (you got to program out outliers, tough, but not impossible), it works only with Q like stocks that have sufficient activity in 5 minutes windows. When you do that, you can see that a channel up is a step function, and with the Q, often a step will march few equivalent boxes at a time. Furthermore, the boxes have "memory" of about two weeks back (there I use hourly for the boxes), giving you area of support and resistance for day trading forays. I hope it helps. Analyze fe of the Q, and you will see that in ascending markets the boxes are actually slanted (almost like a parallelepiped). Don't wait to reenter at the bottom of the box, always raise your reentry point.... The inverse in down markets.
Market Data 
Markets 
AI
