Good systems, bad targets!

[GLC]

Fellow Members.

I want to pose another perspective than the one I have operated on previously.

I have thought that the larger the bank roll and the higher the stop loss, the better.  I used the logic that you should pick the largest stop loss you could afford and hope that you can get ahead of that point and stay ahead.  With a little luck, you might be up for the rest of your life.

Most of our better systems start out doing well, but at some point they have a really bad streak that gets them so deep, it's almost impossible to work your way back out of the hole without risking larger and larger bets.  If these larger bets are lost, we are sunk.

What if we set a reasonable win target, lets say 10 units and we also set a modest stop loss, let's say 10 units or maybe 15 but at most 20.  We would have short sessions, but by stopping early on a down swing, even though some of the downswings would have recovered and reached our win target, we would never lose more that a small bank. 

In this way, we would know on a daily basis how the system is doing.  We could play multiple mini-sessions.  Since we have a pretty good system that usually wins, it's just the occasional down swing that eventually kills us, maybe we can stay ahead and stop often enough to come out on top.

This idea keeps us not too far removed from flat betting since we don't need as large a bank roll as if we were using a 10 step martingale fro even chances or a 4 step martingale for 2:1 bets. 

In reviewing some of my records, I see many times that I would have reached +10 many more times than I would have lost -15.  If we assumed that when we get to -15 we are having a bad session, even though some may turn around, we will surely save ourselves from reaching that point of no return.

The thing that will kill this idea is if we are hitting our stop loss too many times relative to the number of times we are hitting our win target.  This would be a bad system, maybe.

All thoughts welcome.

Thanks,

George

[F_LAT_INO]

George my friend.
Apsolutely agree with you and its evident in one part
of my CWB that I implement it at one stage--short sessions win or lose-
but ppl.aren't patient,testing something over 100000 spins,instead of
splitting it in short day or night sessions or even better as you have suggested.

[mr.ore]

It can't work for too long, maybe if you increase unit size as you win sessions, and decrease stoploss as time goes, then if you get winning sessions streak, it might work as there long enough winning streaks between losing sessions. We need bet selection, that assures, that after a losing streak, there would not be another one too soon. Some zig-zag where zigs are long enough, even if zags are longer because of HE.


somethink like this:

\
\    /\
  \  /  \    /\
   \/    \  /  \    /\
          \/    \  /   \
                 \/      \         
                           \      /\
                            \   /   \
                             \/      \/

but not this:
\
\
  \/\             
      \          / \
       \        /   \   /\
        \/\   /      \/   \   /\
            \/              \/   \
                                   \/

we cannot suppose there would be winning streaks (positive expectation), because they might not be there, so we need a progression. It is ok to use it as long as there are long enough streaks where it works, then trailing stoploss works for avoiding lose, you go virtual and wait for sht to happen, then you jump back, but there must not come another losing streak, it must survive long enough to make some units and increase bet size (ideally). So - bet selection should be able to avoid concentration of loses/failures.

Small sessions with stoploss might be just another way of trailing stoploss. If after every successful session you decrease stoploss while retaining target and increasing unit size a little, in case of a streak of winning sessions, stoploss may be for example 3 units, while target 10. After losing that session, reset unit size to minimal / or smaller value and continue.

example:

base      target /  stoploss    cumulated   profit after lost
unit                                       win           session
size                                   
---------------------------------------------------
1           10       15                 +10           -15
1           10       15                 +20             -5
2           10       14                 +40             -8
2           10       12                 +60           +16
2           10        5                  +80           +50
3           10        3                +110            80-3*3=71

Just an idea...

[Proofreaders2000]

"What if we set a reasonable win target, lets say 10 units and we also set a modest stop-loss, let's say 10 units or maybe 15 but at most 20.  We would have short sessions, but by stopping early on a down swing, even though some of the downswings would have recovered and reached our win target, we would never lose more that a small bank."-GLC

Ten units profit may be too high even then.  How about multiple sessions of +2 units. Maybe a stop-loss of -10.  It's good to remember that even a 'reasonable' target maybe too much, even for a reliable system sometimes.

[mr.ore]

In roulette, chance to win is at best (with one hit wins it method):

p = (36/37) * stoploss/(stoploss+target)

with any method, it is

p = k * (36/37) * stoploss/(stoploss+target)

where  0 < k <= 1, not all methods are same good (because they does not win in one hit)

the greater stoploss, the greater chance we reach target with that particular method.

We need a HUGE bankroll, or A LOT OF luck, or a MAGIC DEVICE, like an ORACLE