StockFetcher Forums · Filter Exchange · TRO'S CROCK POT 2.0<< 1 2 3 4 5 ... 8 >>Post Follow-up
Kevin_in_GA
4,552 posts
msg #115856
Ignore Kevin_in_GA
10/13/2013 2:47:05 PM

This is the start of a parallel conversation regarding TRO's Crock Pot strategy found here:


http://forums.stockfetcher.com/forums/Filter-Exchange/TRO-CROCK-POT-SLOW-COOKERS-FOR-SWING-TRADERS/67200


The basic premise is that one can use historical likelihoods of reaching a 1% or 2% gain from a trade entered at the open of the week's trading. If one places a limit order at 1-2% above the weekly open one can usually meet this goal with little or no effort.


TRO's original filter calculated the frequency of hitting a given % gain from the weekly open, and advocates taking positions in those stocks with the highest win frequency. In his original version, the close must be above 1 dollar and the average volume(90) be greater than 500000.

What was not considered was that when these gains were made, the price and volume criteria must also be met, or one is comparing apples and oranges - how confident are you in meeting a 2% gain this week on a $5 using historical data from when it was trading at $1 or $2?

SF can correct for this however, and so I did a small series of modifications to TRO's original filter to also insure that only trades which met the entry requirements were counted in the statistical analysis.

Fetcher[

/* TRO STAT SCAN 2.0 - use only on Saturday and Sunday */

set{whiop, weekly high - weekly open}
set{Long_Profit, whiop / weekly open }
set{wkProfitPct, 100 * Long_Profit }

set{weeklyopen, count(weekly open above 5,1)}
set{weeklyvolume, count(weekly volume above 2500000,1)}
set{twopcthit, count(Long_Profit > .02 , 1)}

set{tradeentry, weeklyopen * weeklyvolume}
set{tradeentry52, count(tradeentry above 0.5,52)}
set{profithit, tradeentry * twopcthit}
set{twopct52, count(profithit > 0.5 , 52)}
set{winpct52a, twopct52 / tradeentry52}
set{winpct52, winpct52a * 100}

set{tradeentry13, count(tradeentry above 0.5,13)}
set{twopct13, count(profithit > 0.5 , 13)}
set{winpct13a, twopct13 / tradeentry13}
set{winpct13, winpct13a * 100}

add column wkProfitPct
add column tradeentry52 {entries over the last 52 weeks}
add column twopct52 {hit 2% target over last 52 weeks}
add column winpct52 {52 week win percentage}
add column separator
add column tradeentry13 {entries over the last 13 weeks}
add column twopct13 {hit 2% target over last 13 weeks}
add column winpct13 {13 week win percentage}
add column separator
add column weekly open {wopen}
add column weekly high {whigh}
add column weekly low {wlow}
add column weekly close {wclose}
add column separator

close above 5
twopct52 above 40

draw twopcthit
sort column 7 descending

chart-display is weekly
]



This version only includes trades where the weekly open was above 5 and the weekly volume was above 2.5 M shares (similar to TRO's average daily volume of at least 500000).

How different are the results? In some cases not at all, but in others it is pretty dramatic. As an example look at HSOL - on TRO's original filter it shows 49 successful 2% gain weeks over the last 52, but almost all of those were when it's price was below $5. Using this variant, you get a score of 4. Now the filter does not say "Gosh, if you invest in HSOL you won't make your target", but it does not mislead you with results that were from when it traded at a much lower price.

In fact, for the coming week only three stocks are common to both filter's 10 top rank - GMCR, DUST, and TSLA.

I also built in a second time scale - 13 weeks - since for some traders they more recent data might be seen as more important than giving weight to wins made almost a year ago. One just needs to change the selection criteria as follows:


Fetcher[
/* TRO STAT SCAN 2.0 - use only on Saturday and Sunday */

set{whiop, weekly high - weekly open}
set{Long_Profit, whiop / weekly open }
set{wkProfitPct, 100 * Long_Profit }

set{weeklyopen, count(weekly open above 5,1)}
set{weeklyvolume, count(weekly volume above 2500000,1)}
set{twopcthit, count(Long_Profit > .02 , 1)}

set{tradeentry, weeklyopen * weeklyvolume}
set{tradeentry52, count(tradeentry above 0.5,52)}
set{profithit, tradeentry * twopcthit}
set{twopct52, count(profithit > 0.5 , 52)}
set{winpct52a, twopct52 / tradeentry52}
set{winpct52, winpct52a * 100}

set{tradeentry13, count(tradeentry above 0.5,13)}
set{twopct13, count(profithit > 0.5 , 13)}
set{winpct13a, twopct13 / tradeentry13}
set{winpct13, winpct13a * 100}

add column wkProfitPct
add column tradeentry52 {entries over the last 52 weeks}
add column twopct52 {hit 2% target over last 52 weeks}
add column winpct52 {52 week win percentage}
add column separator
add column tradeentry13 {entries over the last 13 weeks}
add column twopct13 {hit 2% target over last 13 weeks}
add column winpct13 {13 week win percentage}
add column separator
add column weekly open {wopen}
add column weekly high {whigh}
add column weekly low {wlow}
add column weekly close {wclose}
add column separator

close above 5
twopct13 above 11

draw twopcthit
sort column 12 descending

chart-display is weekly

]



I believe that this approach can be easily implemented and trade profitably. I will track how the top 5 picks from each time frame do each week and post them here since this is tricky to backtest properly. If I get the time I'll manually backtest this strategy over the last year and post the result as well.


Not sure if a stop loss needs to be put in place here, since some of the losses can be big, wiping out several weeks of gains in one bad trade. I guess we'll see going forward.


klynn55
425 posts
msg #115859
Ignore klynn55
10/13/2013 4:20:01 PM

kevin thanks for being so thoughfull

mahkoh
894 posts
msg #115864
Ignore mahkoh
10/13/2013 6:32:38 PM

I would suggest to always check non adjusted historical data on candidates from a source like Yahoo Finance.

As an example USU (which is not on the list) looks to have traded between $7 and $16 in SF data. But prior to the 1 for 25 split in July 2013 it actually traded below $ 0.70.

And with a few formulas in excel you can check whether your GTC order would have been filled within 3 weeks if it failed to do so during the week. The stats for the top 10:

GMCR: 5 failures, closed out 3 times profitable within 3 weeks ( failure from week ending 9/27 has a chance to close this week)
SRPT : 7 failures, 3 closed out profitable
TSLA : 6 failures, all closed out profitable
BBRY : 8 failures, 6 closed out profitable
YOKU : 8 failures, 3 closed out profitable
P : 8 failures, 7 closed out. This weeks failure still has 3 weeks to try and close with a profit.
UVXY : 8 failures, 2 closed out profitable
DUST : 4 failures( out of 46 entries), 3 closed out profitable
NPSP : 9 failures, 2 closed out profitable
QCOR : 9 failures, 5 closed out profitable



Kevin_in_GA
4,552 posts
msg #115865
Ignore Kevin_in_GA
10/13/2013 9:52:42 PM

Mahkoh - good point, especially for low cost stocks that do a reverse split. The normal stock split should not be that much of an issue, but is also worth checking.

Kevin_in_GA
4,552 posts
msg #115866
Ignore Kevin_in_GA
10/13/2013 9:58:24 PM

I'll be tracking the 13 week top performers as we go along here. Based on performance over the past 13 weeks, trades to be entered at the open on Monday October 14th are

SFUN - 100 shares
DUST - 115 shares
CSIQ - 250 shares
YOKU - 170 shares
WLT -340 shares

All have gone 13 for the last 13. Not sure what the open price will be, but these share amounts would put roughly $5000 into each trade.

olathegolf
64 posts
msg #115868
Ignore olathegolf
10/14/2013 2:05:54 AM

Thanks, Kevin for the v2.0. Always good work.

As TRO always said, it's not what you trade but how you trade it.

Undoubtedly, the win percentages are high. I am most interested in the occasion when the 2% is not triggered. In the old TRO post, Frank (Oldsmar52) suggested that he just waits it out - I guess implying that he always wins. My simple backtesting shows otherwise.

So, the big question is exit strategy. Kevin spoke about this in some detail at the end of the old TRO post.

It might also be interesting to see the win% if you (decide to) hold for 2, 3, or 4 weeks.

Kevin_in_GA
4,552 posts
msg #115869
Ignore Kevin_in_GA
10/14/2013 7:23:37 AM

I might argue that with win rates like these the better strategy is to enter more trades, spreading your risk and putting less into each trade. Then the occasional 10-15% whallop is absorbed by the larger number of 2% wins.

jimmyjazz
99 posts
msg #115879
Ignore jimmyjazz
modified
10/14/2013 10:10:53 AM

In the "let it ride" vein, I would argue that it might make sense to let all trades ride to Friday close if they don't hit the exit limit price earlier in the week. Take whatever exit price you get. Many, many times on the Supercharged Crockpot a -10% stop wasn't triggered and exit near Friday close trades roughly balanced out (some up, some down, most near flat).

Last week was the first time I took a beating using Supercharged Crockpot.

FYI, I am in WLT as of near-open this morning, using a long call, and it has gained enough for me to convert to a trailing stop that will ensure I won't lose money (assuming it triggers and executes at that price). No exit criteria to the upside. I'll just let it ride.

oldsmar52
104 posts
msg #115883
Ignore oldsmar52
10/14/2013 11:25:00 AM

In the other TRO thread I mentioned an option if you didn't like the "buy at open" on Monday method. Today was a rather big down opening & I'd said you might want to wait & when the stock went below its Monday open you could put in a buy stop to get in at its original open price. That's what I did today & got in and out of CSIQ, DUST, HSOL, CPRX & STP with 2% on each. That was on Avery's original filter, not the one in this thread. Anyone pay attention to that option? Frank

Kevin_in_GA
4,552 posts
msg #115885
Ignore Kevin_in_GA
modified
10/14/2013 11:30:32 AM

Just a quick note to say that CSIQ, DUST, and WLT have all hit their 2% profit targets. Right now SFUN and YOKU are both profitable ... but the day is young. Could go either way.

UPDATE at 12:39 PM: YOKU also has now hit the 2% target, and SFUN (the only remaining open trade) is up slightly over 1%.

StockFetcher Forums · Filter Exchange · TRO'S CROCK POT 2.0<< 1 2 3 4 5 ... 8 >>Post Follow-up

*** Disclaimer *** StockFetcher.com does not endorse or suggest any of the securities which are returned in any of the searches or filters. They are provided purely for informational and research purposes. StockFetcher.com does not recommend particular securities. StockFetcher.com, Vestyl Software, L.L.C. and involved content providers shall not be liable for any errors or delays in the content, or for any actions taken based on the content.


Copyright 2016 - Vestyl Software L.L.C.Terms of Service | License | Questions or comments? Contact Us
EOD Data sources: DDFPlus & CSI Data Quotes delayed during active market hours. Delay times are at least 15 mins for NASDAQ, 20 mins for NYSE and Amex. Delayed intraday data provided by DDFPlus