"Scaling out" is inferior behavior

[Buy1Sell2]

Quote from romik:

EDITED SCALE OUT VERSION

50 % winning percentage 4 ES Contracts 20 trades 2 pt target 2 pt loss

1st example without scaling out

10 winners 2X(4 Contracts) = $80 pts ($4000)
10 losers 2X(4Contracts) = $80 pts (-$4000)

Net profit 0 before commissions


2nd example with scaling out half at 2 pt

5 winners 2X(4 Contracts) =40 pts ($2000)
2 breakeven X(4 Contracts) =0($0)
1 winner 3X(4 Contracts)=12pts ($600)
1 winner 4X(4 Contracts)=16pts ($800)
1 winner 10X(4 Contracts)=40pts ($2000)
10 losers 2X(4 Contracts) = -80 pts (-$4000)

Net profit before commissions=+$1400

Now run the first example with the higher profit targets that woud be achieved by letting the trade run( I believe you had one trade run to 10 points). Also, remember that we would get out at breakeven on two as well.

 

[ashcroftsinger]

Quote from Buy1Sell2:

1. Position sizing should never be too large for a trader where the stop loss would exceed 2 percent of total liquid net worth. If you want to use smaller positions than that, fine. If you want to adjust your position size with the market on the entry signal that is fine, just don't exceed the 2 percent rule. Whatever the size that you use, the trade should be allowed to run to maturity. You would be better off long term trading only signals that allow you to use the same size or scale into the same size each time (per market).
2. Attractiveness of trades do change, but you rely on your backtesting to tell you what the percentage of winners is. If you see an obvious reversal while the trade is on, you exit fully not partially.
3. The only true edge in trading is capital and the management thereof. Entries and Exits are part of that management.

Here is essentially what my disagreement is about. IMO, position sizing need not be determined ony by personal risk profiles (2% etc.) it could also be determined by how attractive a trade is, what are the current market conditions etc.

This is quite similar to Ed Thorp's position sizing for blackjack strategies. As the deck becomes more favorable, increase your size of the bet, as it deteriorates, decrease your size of the bet, not necessarily in a binary fashion.

 

[romik]

Quote from romik:

EDITED SCALE OUT VERSION

50 % winning percentage 4 ES Contracts 20 trades 2 pt target 2 pt loss

1st example without scaling out

10 winners 2X(4 Contracts) = $80 pts ($4000)
10 losers 2X(4Contracts) = $80 pts (-$4000)

Net profit 0 before commissions


2nd example with scaling out half at 2 pt

5 winners 2X(4 Contracts) =40 pts ($2000)
2 breakeven X(4 Contracts) =0($0)
1 winner 3X(4 Contracts)=12pts ($600)
1 winner 4X(4 Contracts)=16pts ($800)
1 winner 10X(4 Contracts)=40pts ($2000)
10 losers 2X(4 Contracts) = -80 pts (-$4000)

Net profit before commissions=+$1400

What happens after the scale out is unpredictable future average, only past results will indicate true average.

I see m4a1 made a short call in ES J where he is attempting to use a 1:3 using a specific entry strategy, right now there is a reversal signal going on, he is obviously holding, when a scale out trader would have pulled out of 50% position and/or reversed the trade. Is m4a1 to hold this position till he is -2pts down or exits at entry is irrelevant. A scale out trader has made a profit when m4a1 will either make 0 or incur a loss.

 

[thenewguy]

Quote from illiquid:

The emphasis would just be 10% more towards the present trade, as opposed to relying 100% on the past "optimal" target.

The way I look at it, given the way I trade, (look ma no blanket) scaling is done in conjunction with a flexible outlook, as opposed to a closed, rigid one -- systemizing a scaling exit almost defeats the purpose in my opinion. You can argue ad infinitum comparing system A and system B, but that is besides the whole argument for me.

Quote from illiquid:

What you are doing by scaling is placing greater weight on the results of recent trades, while giving less weight to the result of past trades (where the "optimal" exit has been calculated from) -- an analogy would be an EMA vs an MA on "maturity", so to speak.

I misunderstood what you meant in your first post. I agree with this.

TNG

 

[Buy1Sell2]

Quote from ashcroftsinger:

Here is essentially what my disagreement is about. IMO, position sizing need not be determined ony by personal risk profiles (2% etc.) it could also be determined by how attractive a trade is, what are the current market conditions etc.

This is quite similar to Ed Thorp's position sizing for blackjack strategies. As the deck becomes more favorable, increase your size of the bet, as it deteriorates, decrease your size of the bet, not necessarily in a binary fashion.

I'm fine with the adjusting sizes to your tastes as long as the 2 percent rule is not violated. But once in the trade, you stay for the duration unles you see obvious reversal. No sense getting out of a winner.

 

[Buy1Sell2]

Quote from romik:

What happens after the scale out is unpredictable future average, only past results will indicate true average.

I see m4a1 made a short call in ES J where he is attempting to use a 1:3 using a specific entry strategy, right now there is a reversal signal going on, he is obviously holding, when a scale out trader would have pulled out of 50% position and/or reversed the trade. Is m4a1 to hold this position till he is -2pts down or exits at entry is irrelevant. A scale out trader has made a profit when m4a1 will either make 0 or incur a loss.

If there is an obvious reversal signal , you get out of the full position not part.

 

[thenewguy]

Quote from Buy1Sell2:

I'm fine with the adjusting sizes to your tastes as long as the 2 percent rule is not violated. But once in the trade, you stay for the duration unles you see obvious reversal. No sense getting out of a winner.

According to your argument, why wouldn't you take the largest statistical move the instrument has ever made and use that as a price target? Your math is correct in that if the price target actually has ANY chance of occuring at all, it's best to hold full size to that. How do you know that ahead of time however?

TNG

 

[romik]

Quote from Buy1Sell2:

If there is an obvious reversal signal , you get out of the full position not part.

same as with a scaled out trade

 

[Buy1Sell2]

Quote from thenewguy:

According to your argument, why wouldn't you take the largest statistical move the instrument has ever made and use that as a price target? Your math is correct in that if the price target actually has ANY chance of occuring at all, it's best to hold full size to that. How do you know that ahead of time however?

TNG

If that was your price target and the expectancy was sufficient, it would make sense. If the expectancy is better with overall profit at a lesser price, then that would make sense. Each trader must do this homework .

 

[Buy1Sell2]

Quote from romik:

same as with a scaled out trade

Getting out of the full position is not the same as scaling out. If there was an obvious reversal signal why would you leave some on?