NexusFi: Find Your Edge


Home Menu

 





All in all out vs. scaling in and out


Discussion in Psychology and Money Management

Updated
      Top Posters
    1. looks_one Itchymoku with 11 posts (8 thanks)
    2. looks_two PeakGrowth with 10 posts (26 thanks)
    3. looks_3 Fat Tails with 10 posts (68 thanks)
    4. looks_4 xelaar with 8 posts (4 thanks)
      Best Posters
    1. looks_one sixtyseven with 9.5 thanks per post
    2. looks_two Fat Tails with 6.8 thanks per post
    3. looks_3 Jigsaw Trading with 4.8 thanks per post
    4. looks_4 PeakGrowth with 2.6 thanks per post
    1. trending_up 54,527 views
    2. thumb_up 274 thanks given
    3. group 55 followers
    1. forum 130 posts
    2. attach_file 9 attachments




View Poll Results: All in all out or scale in and out?
All in, all out 87 32.34%
All in, all out
87 32.34%
All in, scale out 115 42.75%
All in, scale out
115 42.75%
Scale in, scale out 67 24.91%
Scale in, scale out
67 24.91%
Voters: 269. You may not vote on this poll

 
Search this Thread

All in all out vs. scaling in and out

  #101 (permalink)
 timendaGain 
New York, NY
 
Experience: Beginner
Platform: NinjaTrader
Broker: NinjaTrader
Trading: ES MES
Posts: 792 since Nov 2010
Thanks Given: 149
Thanks Received: 235


sixtyseven View Post
@timendaGain,
Adding - and exposing yourself to more potential risk - also exposes yourself to more potential gain.

I’m not an expert on this so bear with the density of my skull….
When 5 contracts are added to an initial position of 10 one is positioned with 15 contracts. While there may be some value in comparing that trade of 15 contracts with the AIAO system trading of 10 contracts, I think it’s important to also compare full size positions with one another - 15 Add to 15 AIAO. In doing so I see the Add method as a safer version of the AIAO while AIAO yields a higher return. I modified your spreadsheet to show this. In all cases profitability begins at a 51% edge, as one would expect.

While putting on a trade it feels safer to start with a smaller size. For reversion to the mean styles of trading, adding when price goes against you can help you make individual trades work out.

I think that the reason why adding/scaling in is advised is because it's too difficult to get the initial entry just right consistently enough. Instead we need to spread our entries across multiple price points to get more even trading results. This means that sometimes the initial entry is right and the trader completes the trade with a smaller size.

I used the example of scaling in when price moves against you. I fully understand that style of increasing the position size.

I'm struggling with why I would add more when it's moving in my favor. I just don't think that the market is more likely to continue moving in my favor just because it has been. In fact, I find it less likely.

I don’t understand the difference between scaling and adding on your spreadsheet. Please explain.

Attached Files
Elite Membership required to download: Risk and scale in.xlsx
Visit my NexusFi Trade Journal Reply With Quote
Thanked by:

Can you help answer these questions
from other members on NexusFi?
Are there any eval firms that allow you to sink to your …
Traders Hideout
Deepmoney LLM
Elite Quantitative GenAI/LLM
NexusFi Journal Challenge - April 2024
Feedback and Announcements
NT7 Indicator Script Troubleshooting - Camarilla Pivots
NinjaTrader
Exit Strategy
NinjaTrader
 
Best Threads (Most Thanked)
in the last 7 days on NexusFi
Get funded firms 2023/2024 - Any recommendations or word …
61 thanks
Funded Trader platforms
39 thanks
NexusFi site changelog and issues/problem reporting
26 thanks
The Program
18 thanks
GFIs1 1 DAX trade per day journal
18 thanks
  #102 (permalink)
 Itchymoku 
Philadelphia
 
Experience: None
Platform: corded black telephone
Trading: ticker tape
Posts: 2,894 since Apr 2012
Thanks Given: 1,683
Thanks Received: 3,681

Basically the market doesn't just go in one straight line up or down, so why would it make any sense to trade it in such a static fashion. That's the conclusion I've come to with testing not only all in/all out vs scaling in/scaling out, but also with scalping considering it runs along the same concept (price is more chaotic on a lower level). Trading a little further out generally equates to price moving from point a to b with more stability. I've found being able to scale into a position particularly helpful too.

R.I.P. Joseph Bach (Itchymoku), 1987-2018.
Please visit this thread for more information.
Reply With Quote
Thanked by:
  #103 (permalink)
 Itchymoku 
Philadelphia
 
Experience: None
Platform: corded black telephone
Trading: ticker tape
Posts: 2,894 since Apr 2012
Thanks Given: 1,683
Thanks Received: 3,681



Silvester17 View Post
just out of curiosity a few questions

scaling in:

- the majority of my losing trades, I'm fully loaded. the majority of my winning trades, I'm not

scaling out:

- I only scale out of winning trades, but I don't scale out of losing trades. (the answer to get out of a losing trade because it's not valid anymore can also be used for a target)

Are they questions for me?



if you scale into a winning trade it should be far enough in your favor that if it comes back to break even it's a good idea to get out or scale out at break even. I try to scale in when the market pulls back after going in my favor. Try to hold all of your scales till the trade hits your limit order. (it's probably not a good idea to scale in too much even if the trade is going your direction unless you have a enough cushion room to your net break even where you can scale out*)

If the trade is going against you I'd only suggest to scale in once, not repetitively. If then the trade comes back to break even or you recoup some of your loss it's okay to scale out and minimize your risk. Then if the trade goes in your favor past your break even mark and gives you some wiggle room it's okay to scale in again like in the first paragraph. I hope that clears up some confusion. That's what works best for me for the most part.

R.I.P. Joseph Bach (Itchymoku), 1987-2018.
Please visit this thread for more information.
Reply With Quote
Thanked by:
  #104 (permalink)
 PeakGrowth 
Sydney, Australia
 
Experience: Intermediate
Platform: Sierra Chart, IRESS
Broker: IB, IQFeed
Trading: ES, SPI, ASX stocks, options
Posts: 399 since Jun 2015
Thanks Given: 169
Thanks Received: 465

A couple of points.

Scaling out will never produce a higher expectancy than all out if you have back tested your system and know your optimal exit, this is a mathematical fact. If you have 2 scale out points (whether it be a fixed + runner, or whatever), it will always be the average of the two. If through back testing, one scale out has a PF of 1.5 and one has a PF of 2.0, then scaling out will produce 1.75, which will always be lower than the most optimal exit at 2.0. What this means is that you are theoretically trading the average of two systems, therefore, if you want to scale out, you should back test both systems to ensure they both have a positive PF, otherwise, you will be trading sub optimally for no real reason.

Scaling out will smooth out your equity curve but at the expense of expectancy. This counteracts each other so the higher leverage allowed due to the smoother equity will only bring your return back to approximately where your previous return would have being, defeating the purpose. It will actually be cheaper due to commissions and slippage to just trade the higher expectancy one with lower leverage and higher draw down.

This basically means that scaling out won't add any value to a system which does not already have a positive expectancy and you should be careful not to be deluded that you can turn a losing system into a winning one just by scaling out.

Additionally, the argument that you don't know what your optimal exit is, is not true. If you cannot pick an exit point with added expectancy, you should just use a time exit or do more research. Your entry and exit should both add expectancy, not take away from it. For example, if you test a system with a 2 pt stop loss, 1 point exit (some people scale out at this point to pay for the trade) and a 4 point exit with the same stop loss, it is quite obvious through back testing the 1 point exit in almost all cases simply does not add expectancy and hence, should not be scaled out at this point. The 4 point is optimal and the 1 point is not. If you don't know what your expectancy is for each scale out - you should probably find out.

Finally, scaling out is good psychologically for the discretionary trader, but does not add value mathematically. There is nothing wrong with scaling out if it helps you psychologically.

Scaling in is a very good strategy, depending on your market and time frame and is very situational. For scaling in to work, you need to be in a market that trends strongly and used on a time frame higher than intraday. If you are a momentum trader on the intraday time frame, scaling in doesn't work because trends don't last long enough for scaling in to work. This is especially true for ES traders. Daily and above is where scaling in works as the trends can last longer as there is no time constraint like intraday traders. Additionally, scaling in should not be used with targets (unless you can pick them with a very high expectancy) as large and uncommon trends need to be allowed to run to pay for all the failed scale ins.

Scaling in causes a much more volatile equity curve and will flip most systems with a higher than 50% win rate to lower than 50% and further increase your win/loss ratio. The main issue with this is that it increases draw down periods and losses in a row significantly, so scaling in is counter productive for the discretionary trader psychologically, but increases PF in general if done properly at the right time, on the right market, on the right time frame.

In terms of what is the right time to scale? This can only be answered either on a discretionary basis or on a back tested basis. If you know the runup distribution for your losing trades and scale in outside of the first/second or third SD of this distribution, then scaling in will add PF. If your losing trade runup distribution is too wide, it means your entry is not good enough or your exit is not good enough and you should not scale in. It is also important you know the distribution of your winning trade run ups so that you know the trends you are capturing have enough gas in them for scaling in to work. Scaling in is extremely hard to do well, if you don't have an established edge you will only exacerbate how quickly you'll blow yourself up. Hence if you are new to trading, don't do it.

The conclusion is this:
- Scaling out helps psychologically but reduces PF and blows you up slower if your system is not positive expectancy.
- Scaling in damages psychologically but increases PF and blows you up much quicker if your system is not positive.

If you want the highest expectancy, you should scale in all out on daily+ timeframes. All in all out for intraday.
If you want the least psychologically damaging method that allows you to stay in the game for longer so you can help yourself beat yourself, then you should go all in and scale out.

Visit my NexusFi Trade Journal Reply With Quote
  #105 (permalink)
 Itchymoku 
Philadelphia
 
Experience: None
Platform: corded black telephone
Trading: ticker tape
Posts: 2,894 since Apr 2012
Thanks Given: 1,683
Thanks Received: 3,681


PeakGrowth View Post
Scaling out will never produce a higher expectancy than all out if you have back tested your system and know your optimal exit, this is a mathematical fact.

There's no way to mathematically know the outcome of a trade based on previous data. The information within a trade is much more relevant than information before the trade has started. Scaling allows you more options to choose which data points have out preformed and which have under preformed, whilst using a single entry or exit throws all of your available resources to the whim of the market.

Scaling into a position doesn't necessarily have to be used on a trending market either. It does have some psychological benefits, but that's not the reason I like to use it.

I've rarely ever seen anyone use a system that works long term that uses all in - all out. If you do, let me know.

R.I.P. Joseph Bach (Itchymoku), 1987-2018.
Please visit this thread for more information.
Attached Thumbnails
Click image for larger version

Name:	Fibonacci.jpg
Views:	258
Size:	88.1 KB
ID:	200068  
Reply With Quote
Thanked by:
  #106 (permalink)
 Itchymoku 
Philadelphia
 
Experience: None
Platform: corded black telephone
Trading: ticker tape
Posts: 2,894 since Apr 2012
Thanks Given: 1,683
Thanks Received: 3,681

Actually I'll even go as far to say that back-testing doesn't really work either. You're basically just exiting on arbitrary figures. It may help to back-test, but I think there's more required to know where to exit than just averaging mfe on previous trades that meet some set of criteria. When you start doing more research you're basically just curve-fitting. There's no end to it.

R.I.P. Joseph Bach (Itchymoku), 1987-2018.
Please visit this thread for more information.
Reply With Quote
Thanked by:
  #107 (permalink)
 PeakGrowth 
Sydney, Australia
 
Experience: Intermediate
Platform: Sierra Chart, IRESS
Broker: IB, IQFeed
Trading: ES, SPI, ASX stocks, options
Posts: 399 since Jun 2015
Thanks Given: 169
Thanks Received: 465


Itchymoku View Post
Actually I'll even go as far to say that back-testing doesn't really work either. You're basically just exiting on arbitrary figures. It may help to back-test, but I think there's more required to know where to exit than just averaging mfe on previous trades that meet some set of criteria. When you start doing more research you're basically just curve-fitting. There's no end to it.

U are making the assumption that an optimal exit is based on averaging an mfe, which reduces expectancy. I never said that was the case. Optimal can be achieved in a couple of ways, for one, knowing that you don't know what optimal is and using a time exit such as EOD can work. There is an optimal and there is at least ones that are sub optimal. For example, a 2 tick vs a 10 tick target will have different results and one will most definitely be better than the other. By knowing this you can at least avoid a strategy that just won't work at all.

Sorry but back testing does work, if your system can't survive a back test then it will never survive in real trading. A back test is simply used to prove that your hypothesis of a market efficieny can survive in the past so there is a larger chance it will survive in the future. Back testing is used to prove survivability and an inefficieny in the past not predict returns.

I would go as far as to say if you don't back test your idea you are throwing your money away.

By looking for patterns either with discretion or not you are by definition curve fitting, the issue with developing a method is not curve fitting, it is over fitting with too many rules.

Sent from my SM-N9005 using Tapatalk

Visit my NexusFi Trade Journal Reply With Quote
Thanked by:
  #108 (permalink)
 Itchymoku 
Philadelphia
 
Experience: None
Platform: corded black telephone
Trading: ticker tape
Posts: 2,894 since Apr 2012
Thanks Given: 1,683
Thanks Received: 3,681


PeakGrowth View Post
U are making the assumption that an optimal exit is based on averaging an mfe, which reduces expectancy. I never said that was the case. Optimal can be achieved in a couple of ways, for one, knowing that you don't know what optimal is and using a time exit such as EOD can work. There is an optimal and there is at least ones that are sub optimal. For example, a 2 tick vs a 10 tick target will have different results and one will most definitely be better than the other. By knowing this you can at least avoid a strategy that just won't work at all.

Sorry but back testing does work, if your system can't survive a back test then it will never survive in real trading. A back test is simply used to prove that your hypothesis of a market efficieny can survive in the past so there is a larger chance it will survive in the future. Back testing is used to prove survivability and an inefficieny in the past not predict returns.

I would go as far as to say if you don't back test your idea you are throwing your money away.

By looking for patterns either with discretion or not you are by definition curve fitting, the issue with developing a method is not curve fitting, it is over fitting with too many rules.

Sent from my SM-N9005 using Tapatalk

Yeah, but you'll always be back testing on data that has already happened even if you are comprehensive with applying it to all different sorts of periods of volatility or types of market behavior. By choosing the data you pick, or even not choosing it, you are assuming the market will act similar going forward live as the market behaved previously.

R.I.P. Joseph Bach (Itchymoku), 1987-2018.
Please visit this thread for more information.
Reply With Quote
  #109 (permalink)
 
Tymbeline's Avatar
 Tymbeline 
Leeds UK
Market Wizard
 
Experience: Intermediate
Platform: Tradovate
Broker: Tradovate
Trading: MES, MNQ
Frequency: Several times daily
Duration: Minutes
Posts: 644 since Apr 2015
Thanks Given: 2,342
Thanks Received: 1,054

As a member who joined this year, I'm glad this old (but timeless) thread was bumped, today, because I hadn't read it before and found it particularly interesting.

Having now read it all, two main things strike me, here.

First, there's a kind of widespread background assumption (actually it's also been expressly stated in a couple of places) that "scaling is of no value to scalping". My take-home from this (especially given that Big Mike, who describes himself as a scalper, is routinely using scaling - and I'm sure he's far from the only steadily profitable, professional trader in that position) is that much depends on one's exact definition/description of what's classified as "scalping": it's one of those terms rarely precisely defined, and therefore rather prone to being discussed somewhat at cross-purposes.

Secondly:-

baruchs View Post
Scaling in/out is logically wrong.
Maybe psychologically its OK.

This is undoubtedly far more true for some than for others.

Personally, I find most psychological benefit in trading from the things that satisfy me logically and evidentially in the statistical/probabilistic sense, because that way they become part of my edge-awareness confidence.

Reply With Quote
Thanked by:
  #110 (permalink)
 
Jigsaw Trading's Avatar
 Jigsaw Trading  Jigsaw Trading is an official Site Sponsor
 
Posts: 2,988 since Nov 2010
Thanks Given: 831
Thanks Received: 10,393



PeakGrowth View Post
A couple of points.

Scaling out will never produce a higher expectancy than all out if you have back tested your system and know your optimal exit, this is a mathematical fact. .

Your fact is only a fact if you believe in systems. I too was once a believer....

A system being something mechanical with a defined entry and exit point.

But the fact is - to agree with your argument, you have to take a leap of faith that a mechanical system is possible in the first place. So your fact is based on a leap of faith.

Discretionary directional trading is based on right here, right now, the market is directional (or non-directional) and that the final outcome cannot be known but that scaling out is one way to ride the trade out.

All this talk of back testing, expectations, systems, ratios has little relationship to real world proprietary trading. Nor does it bear any resemblance to algorithmic trading. It's really the realm of a number of retail 'automation' platforms - where there's scant evidence that a single trader has ever automated their way to profit.

To put some perspective on this - consider the hundreds of millions of dollars invested by HFT firms to be the fastest link between disparate servers in order to scalp a penny of price discrepancy. Would they really invest that money to arb latency of they could just write a "strategy" to directionally trade the market?

I don't think so and I am yet to hear a convincing argument as to why automated directional (non-arb) trading is feasible.

Over the past few months there have been times when the market has become more volatile and at some points liquidity took a huge drop. On those days, you are more likely to see runners, so scaling in makes sense. In the summer time when the range traditionally drops, all in-all out makes more sense. On normal days, I can see an argument for scaling out because that is what I do on those days.

So for me it's like this. I let the type of action dictate the way I enter AND the way I manage a trade. There's no way I'm going all in when my chosen market is slipping 6 ticks one way or the other in a few seconds. But when it slows down to a crawl, how can I realistically do anything other than put on the max position and get out at a fair profit when the chance of a runner is close to zero.

Just know your market. Know that exit isn't a one size fits all and understand that "trading = mathematical" is just one of many religions in trading that may or may not be true.

If you have any questions about the products or services provided, please send me a Private Message or use the futures.io " Ask Me Anything" thread
Visit my NexusFi Trade Journal Reply With Quote




Last Updated on December 27, 2015


© 2024 NexusFi™, s.a., All Rights Reserved.
Av Ricardo J. Alfaro, Century Tower, Panama City, Panama, Ph: +507 833-9432 (Panama and Intl), +1 888-312-3001 (USA and Canada)
All information is for educational use only and is not investment advice. There is a substantial risk of loss in trading commodity futures, stocks, options and foreign exchange products. Past performance is not indicative of future results.
About Us - Contact Us - Site Rules, Acceptable Use, and Terms and Conditions - Privacy Policy - Downloads - Top
no new posts