Post New

Somewhere back in the past I discovered that with volatile screens such as Gentle Screamers you could improve overall returns by selling a position if it shot up at least 30%. There was a good chance the price would recede, and it was better to capture the spike. Hence, a profit latch.

Recently I found myself wondering if the concept would work on less volatile screens. Not at the 30% level, obviously, but rather at a more sane 5%. The screen I chose to look at was Advanced because I knew nothing about it (no bias) and the screen picks had small spreads (on the order of 0.1% or less).

Using the top five picks with a 4-week hold, I tracked the returns beginning in 12/30/21, using the opening prices on the following Monday and assumed good-until-canceled sell orders were entered just after the purchases. The sell orders were 5% above the opening price.

The three years showed the following (compounded) returns:

             Basic     with 5%
      IWM    Screen    latch
2022 -20%    - 9%       25%
2023  17%    - 5%       13%
2024  11%      3%       14%

A promising beginning.

DB2

Wow talk about post from the past. This was one of my few fruitful discussion I started way back when Zee was discussing Decision Moose and his Ride the Wave strategy. It was a very long and productive discussion.

There was two different thoughts about implementing a profit latch with the first being once a holding reaches X% (15-25% I believe was the optimal "mound of toast"), you sell the whole lot, be happy with that gain and wait for the next signal. The other thought was when you reach you X%, you cash out a a certain % of your holding and continue holding the rest in hope for further gains. Reason behind this is there was several cases where you continued to hold the same equity for several months and missed out on another big 30% move while sitting on sideline waiting for a new signal.

I personally have adopted locking in a profit of a certain portion of my holding and let the rest run till it hits another profit latch or falls off the screen. Overall I think this has helped my portfolio return & volatility by locking in profits, having a cash cushion on the sidelines all while still keeping me in the game. However other single outliers like NVDA & ANET it has hurt by not holding onto my full initial investment for the last 2 years.

Here is the original link to the thread, however it is no longer working. If anyone has an archive of the old MF MI pages it would be great if we could revive it for further backtesting and discussion.

https://discussion.fool.com/profit-latch-with-scre...

Peace,
Opihi

Hi DrBob,

This is very interesting, and there are some details I wanted to clarify:

(1) What do you do if you stop-gain out of a stock, but it remains on the screen the next month? Do you reinvest in it, or keep it in the "stopped out" pile until something else replaces it?

(2) Did you backtest this on GTR1, and if so, how did you do so (or if anyone else can help me do this on GTR1, I'd love to try).

(3) Is there a "mound of toast" involved? Did you seek out which stop-gain level maximized returns or minimized drawdowns or any other interesting metric to optimize.

There are a few screens I would like to try this with, to see if they might add some value.

Thanks, Gabriel

What do you do if you stop-gain out of a stock, but it remains on the screen the next month?

For this test I just took the list of tickers every four weeks, so if a stock was still on the list it was "reused".

Did you backtest this on GTR1

No, I'm not any good with GTR1, so it was done by hand.

Is there a "mound of toast" involved?

Since it was done by hand, I only tested the one (relatively low) profit latch. Consider it a proof of concept or extension of concept, if you will.

DB2

OK, following up on this to the rest of the board, is there anyone who can help me with the GTR1 to implement a profit latch. I'd like to do it both ways, with a "HTD" that is you don't reinvest until the stock has fallen off the screen and come back on, and also with a reuse (that is, if you have stopped out of the stock but it comes back on the screen, you rebuy it).

I'd also like to be able to look at the mound of toast and see how it changes with the various profit latche percentages, and see if that surface is smooth or jagged.

Thanks, G.

Is there a "mound of toast" involved?

Here is Mr. Toast's mound:

Profit
Latch   '22    '23    '24   Total
  3%   1.083  1.029  1.068   19%
  5%   1.254  1.135  1.138   62
  7%   1.353  1.164  1.108   74
  9%   1.457  0.930  1.073   45
 11%   1.447  0.912  1.110   46
.
 13%   1.407  0.914  1.035   33
 15%   1.391  0.913  1.077   36
 17%   1.352  0.920  1.106   37
 19%   1.190  0.933  1.005   12

DB2

Adding a 5% stop gain to {Advanced} leads to lower CAGR and lower GSD. GTR1 results from 19870302 to 20250324 (5 deep, no friction, 20 day hold):

      Screen         CAGR  GSD  Sharpe
      Advanced        32   34    1.00
Advanced_05stopgain   21   21    0.97

The 5% stop gain led to higher Total Return in recent years:

YearEnding  Advanced  Advanced_05stopgain
 20221230     -14             -3
 20231229      -2             10
 20241231      19             18

A stop gain can improve some screen results. For example, the 20% stop gain in {NetNetsNCAV21} greatly improves results. GTR1 results from 19870302 to 20250324 (10 deep, 2% friction, 30 month hold):

         Screen           CAGR  GSD  Sharpe
      NetNetsNCAV21        32   50    0.76
NetNetsNCAV21_NoStopGain    3   74    0.27

Another example. Adding a 20% stop gain to {NetNetsGrahamified} improves results. GTR1 results from 19870302 to 20250324 (5 deep, 2% friction, 21 day hold):

            Screen             CAGR  GSD  Sharpe
      NetNetsGrahamified        15   53    0.43
NetNetsGrahamified_20stopgain   21   44    0.57

https://gtr1.net/2013/?~Advanced::styp.a:et0!10!11...
https://gtr1.net/2013/?~Advanced_05stopgain:g1.05:...
https://gtr1.net/2013/?~NetNetsNCAV21:h21n30r12f2g...
https://gtr1.net/2013/?!!QlpoMTFBWSZTWUg6qYgAAW7fg...
https://gtr1.net/2013/?~NetNetsGrahamified:h21f2::...
https://gtr1.net/2013/?~NetNetsGrahamified_20stopg...

Thanks, lizgdal. Back when I was doing Net-Nets I used a 50% sell point which gives the same CAGR but a higher GSD. Good to know.

I notice the backtester only uses closing prices. Intraday highs would be better.

DB2

Post New

Mechanical Investing FAQ

---

Contact Shrewd'm
Contact the developer of these message boards.