NexusFi: Find Your Edge


Home Menu

 





Tap In's Corner


Discussion in Trading Journals

Updated
      Top Posters
    1. looks_one Tap In with 656 posts (1,161 thanks)
    2. looks_two bobwest with 42 posts (123 thanks)
    3. looks_3 Inletcap with 21 posts (131 thanks)
    4. looks_4 lancelottrader with 20 posts (32 thanks)
      Best Posters
    1. looks_one Inletcap with 6.2 thanks per post
    2. looks_two ratfink with 3.9 thanks per post
    3. looks_3 bobwest with 2.9 thanks per post
    4. looks_4 Tap In with 1.8 thanks per post
    1. trending_up 121,859 views
    2. thumb_up 1,778 thanks given
    3. group 47 followers
    1. forum 855 posts
    2. attach_file 1,600 attachments




 
Search this Thread

Tap In's Corner

  #801 (permalink)
 
Tap In's Avatar
 Tap In 
Bend, OR
 
Experience: Intermediate
Platform: Sierra Charts
Broker: Global Futures/OEC
Trading: NQ, CL
Posts: 948 since Feb 2013
Thanks Given: 2,162
Thanks Received: 2,082


grausch View Post
Hi @Tap In,

I have been doing a fair bit of research on optimum exit points since it is one of the weak areas of my trading. Since you mentioned the difficulty you had with maintaining a 50% win rate while using a 2:1 reward to risk ratio, I thought the following may be of interest to you.

I used various trend identification techniques, namely Donchian Channels, price straddles at certain times (like Donchian Channels, but with price as the mid point), VWAP straddles (like price straddles, but these were only implemented when price was within 5 ticks of VWAP), and some others not worth mentioning. Each day would have two trades, either a long or a short. I did not test counter-trend strategies as I never found any way to really make these work.

Stops were either 20 ticks or 30 ticks - they were mixed up in different test, but once a stop was chosen it was consistently applied for an individual test. For a 20 tick stop profit targets ranged from 2R to 6R. For a 30 tick stop they ranged from 2R to 4R. One exit point was always time-based, i.e. exit at 16h00 unless stopped out before hand.

The results tended to be quite spread out. Most of my tests showed a clear preference for a 20 tick stop - I guess if you're trading breakouts, there just isn't enough profit potential to offset the losses from the 30 tick stops. With a 20 tick stop, the 2R system performed remarkably well with a higher win% and higher profit most of the time. However, when one of the 6R or time-based exits really scored, then the 2R was left behind. 2 large winners a month would swing the results in favour of the longer stops.

Trading 1 contract after 3 months, final win % for the 2R system was around 40% with average monthly profit just over $1k. The 4R system had a win% of 25% with average monthly profit at about $500. The 6R and time based exits had the same win% as the 4R system, but average gains of about $2k and $1.5k respectively.

Moving stops to break-even did not work that well (especially on the 6R and time based exits) since I was stopped out of too many good trades. Pyramiding suffered from the same fate, and if using multiple contracts, taking the full position and scaling out provided much better results - however, when scaling out it was imperative that the high R exits were still being utilised.

To summarise - using mechanical entries that identify trend, a 50% 2:1 ratio may not be optimal and holding for larger gains seems to pay much better.

Interesting results, especially the part about 6R having the same win rate as 4R. It seems like the probabilities would be against this happening unless the sample size was too small.

I find that reaching multiples of R drops off dramatically with higher R. For example, 1R 60%, 2R 28%, 3R 12%, 4R 6%, etc. A system that reaches 6R 25% of the time would be a gold mine! Bear in mind that I do not consider it reasonable for a discretionary trader to hold a trade to 2R, for example, and let it come back to entry without taking some profit along the way. Therefore, I do not count some trades that do this and end up reaching higher multiples of R. Over 90% of my trades that reach 1R and come back to entry ultimately end up hitting my stop with no more favorable movement. I would go nuts watching a 2R trade come back to zero, so I would have to be trading mechanically and walk away to accomplish what you researched.

Thanks for the post

Started this thread Reply With Quote
The following 4 users say Thank You to Tap In for this post:

Can you help answer these questions
from other members on NexusFi?
My NT8 Volume Profile Split by Asian/Euro/Open
NinjaTrader
NexusFi Journal Challenge - April 2024
Feedback and Announcements
Request for MACD with option to use different MAs for fa …
NinjaTrader
ZombieSqueeze
Platforms and Indicators
 
Best Threads (Most Thanked)
in the last 7 days on NexusFi
Retail Trading As An Industry
58 thanks
Battlestations: Show us your trading desks!
55 thanks
NexusFi site changelog and issues/problem reporting
48 thanks
What percentage per day is possible? [Poll]
31 thanks
GFIs1 1 DAX trade per day journal
29 thanks

  #802 (permalink)
 grausch 
Luxembourg, Luxembourg
 
Experience: Advanced
Platform: TWS
Broker: Interactive Brokers
Trading: Stocks
Posts: 494 since May 2012
Thanks Given: 1,731
Thanks Received: 1,159


Tap In View Post
Interesting results, especially the part about 6R having the same win rate as 4R. It seems like the probabilities would be against this happening unless the sample size was too small.

I find that reaching multiples of R drops off dramatically with higher R. For example, 1R 60%, 2R 28%, 3R 12%, 4R 6%, etc. A system that reaches 6R 25% of the time would be a gold mine! Bear in mind that I do not consider it reasonable for a discretionary trader to hold a trade to 2R, for example, and let it come back to entry without taking some profit along the way. Therefore, I do not count some trades that do this and end up reaching higher multiples of R. Over 90% of my trades that reach 1R and come back to entry ultimately end up hitting my stop with no more favorable movement. I would go nuts watching a 2R trade come back to zero, so I would have to be trading mechanically and walk away to accomplish what you researched.

Thanks for the post

You're welcome - I was quite surprised at the results too.

The win rate for 2R was higher than expected. I did not think basing trades on breakouts would lead to that high a win rate and expected a number similar to yours. There seem to be certain times that were more conducive to better trends, and during certain periods, early morning trends (between 2 and 4 am) tended to be very profitable, i.e. over 50% win rate. However, this changed later on and suddenly none of those trades worked. A larger sample may lead to different results, and perhaps I was just lucky with the chosen period.

The 6R system is not really such a goldmine. It gives back a lot of profits and I have had 9 negative days in a row. Sure, average monthly profit was positive, but intra-month swings can be severe. One month the two longer terms systems were up around $4k each and they ended up at $1.5k and 1.8k each - bear in mind this was one of the better performing "systems" for that month.

Overall it was quite disappointing to see results are at best $2k a month using a single contract. Using a $20k account, that would imply risk of 2% per day and returns of 10% per month. Cutting risk to my preferred 0.5% per day, would lead to returns of 2.5% a month. These are also backtested numbers, so it would be wise to place a haircut on them and in that case, the results become even more disappointing. Guess these type of mechanical systems will struggle to beat a truly competent trader.

Reply With Quote
The following 3 users say Thank You to grausch for this post:
  #803 (permalink)
 
Inletcap's Avatar
 Inletcap 
Murrells Inlet SC
 
Experience: Advanced
Platform: Tradestation
Trading: ES, CL, ETFs
Posts: 9,159 since Dec 2012
Thanks Given: 9,765
Thanks Received: 26,036



grausch View Post
Hi @Tap In,

I have been doing a fair bit of research on optimum exit points since it is one of the weak areas of my trading. Since you mentioned the difficulty you had with maintaining a 50% win rate while using a 2:1 reward to risk ratio, I thought the following may be of interest to you.

I used various trend identification techniques, namely Donchian Channels, price straddles at certain times (like Donchian Channels, but with price as the mid point), VWAP straddles (like price straddles, but these were only implemented when price was within 5 ticks of VWAP), and some others not worth mentioning. Each day would have two trades, either a long or a short. I did not test counter-trend strategies as I never found any way to really make these work.

Stops were either 20 ticks or 30 ticks - they were mixed up in different test, but once a stop was chosen it was consistently applied for an individual test. For a 20 tick stop profit targets ranged from 2R to 6R. For a 30 tick stop they ranged from 2R to 4R. One exit point was always time-based, i.e. exit at 16h00 unless stopped out before hand.

The results tended to be quite spread out. Most of my tests showed a clear preference for a 20 tick stop - I guess if you're trading breakouts, there just isn't enough profit potential to offset the losses from the 30 tick stops. With a 20 tick stop, the 2R system performed remarkably well with a higher win% and higher profit most of the time. However, when one of the 6R or time-based exits really scored, then the 2R was left behind. 2 large winners a month would swing the results in favour of the longer stops.

Trading 1 contract after 3 months, final win % for the 2R system was around 40% with average monthly profit just over $1k. The 4R system had a win% of 25% with average monthly profit at about $500. The 6R and time based exits had the same win% as the 4R system, but average gains of about $2k and $1.5k respectively.

Moving stops to break-even did not work that well (especially on the 6R and time based exits) since I was stopped out of too many good trades. Pyramiding suffered from the same fate, and if using multiple contracts, taking the full position and scaling out provided much better results - however, when scaling out it was imperative that the high R exits were still being utilised.

To summarise - using mechanical entries that identify trend, a 50% 2:1 ratio may not be optimal and holding for larger gains seems to pay much better.

Nice work- would it be possible to plug in a second entry near puke point (say 5-10t away). This is going to blow your mind what happens to the $ gain!!!

Visit my NexusFi Trade Journal Reply With Quote
The following 6 users say Thank You to Inletcap for this post:
  #804 (permalink)
 bmtrading9 
Atlanta, GA, USA
Market Wizard
 
Experience: Advanced
Platform: MC and Jigsaw
Trading: ES, MES
Posts: 1,833 since Mar 2013
Thanks Given: 3,001
Thanks Received: 2,159


Inletcap View Post
Nice work- would it be possible to plug in a second entry near puke point (say 5-10t away). This is going to blow your mind what happens to the $ gain!!!

I really like that idea one question though what would be your stop on both contracts? For now I am experimenting with same stop as the original but that makes me nervous bcoz now if I get stop out it will be twice the hole.

Sent from my SAMSUNG-SM-G900A using Tapatalk

Follow me on Twitter Visit my NexusFi Trade Journal Reply With Quote
The following 2 users say Thank You to bmtrading9 for this post:
  #805 (permalink)
 grausch 
Luxembourg, Luxembourg
 
Experience: Advanced
Platform: TWS
Broker: Interactive Brokers
Trading: Stocks
Posts: 494 since May 2012
Thanks Given: 1,731
Thanks Received: 1,159


bmtrading9 View Post
I really like that idea one question though what would be your stop on both contracts? For now I am experimenting with same stop as the original but that makes me nervous bcoz now if I get stop out it will be twice the hole.

Sent from my SAMSUNG-SM-G900A using Tapatalk

My interpretation of puke point is the original stop and I am basing my testing on that. For instance if I entered a position with a 20 tick stop, I am adding another trade halfway to that, i.e. it carries half the risk of the first contract for a total risk of 1.5R.

I am doing these tests manually so that I can compare how the trades work at different points so it might take another day. Preliminary results are very promising, but I know that the dataset has a period of chop coming up and this may change the results.

Reply With Quote
The following 4 users say Thank You to grausch for this post:
  #806 (permalink)
 
Inletcap's Avatar
 Inletcap 
Murrells Inlet SC
 
Experience: Advanced
Platform: Tradestation
Trading: ES, CL, ETFs
Posts: 9,159 since Dec 2012
Thanks Given: 9,765
Thanks Received: 26,036


grausch View Post
My interpretation of puke point is the original stop and I am basing my testing on that. For instance if I entered a position with a 20 tick stop, I am adding another trade halfway to that, i.e. it carries half the risk of the first contract for a total risk of 1.5R.

I am doing these tests manually so that I can compare how the trades work at different points so it might take another day. Preliminary results are very promising, but I know that the dataset has a period of chop coming up and this may change the results.

Thanks and yes- puke point = stop point!

Visit my NexusFi Trade Journal Reply With Quote
The following 2 users say Thank You to Inletcap for this post:
  #807 (permalink)
 grausch 
Luxembourg, Luxembourg
 
Experience: Advanced
Platform: TWS
Broker: Interactive Brokers
Trading: Stocks
Posts: 494 since May 2012
Thanks Given: 1,731
Thanks Received: 1,159

Completed testing now and I am ashamed to admit that my initial profit calculations had an incorrect formula that overstated the results. @Tap In - you were right to be skeptical of the results, they were significantly overstated.

In any case, I have redone the spreadsheets using adding in an easy way to test whether adding on retracements could work. Entries were determined by straddling price at 2am. I have not retested every single system I tried, since it takes quite a while and most of the results were not really that exciting. Trades were held until stops or targets were hit or until 16h00.

This time, the win% is more in line with what I would expect it to be. The best performing systems tend to be those that hold for 6R or more even if the win % was lower.

Since this evolved out of a discussion on pyramiding or averaging down vs a single contract, I will give a summary of the results of using a 20 tick stop and adding once the position has gone 10 ticks against it vs just trading a fixed number of contracts. Results are for the period 1 June 2016 to 31 August 2016 using the CL contract in Ninjatrader (thus I am not sure of how the contract is put together and whether or not it is a continuous contract). Spreadsheet is attached for those who are interested in reviewing the detail. Testing was done manually so there could be some errors, but if there are errors, the impact should be immaterial.

Results with averaging down at the following profit targets as follows:
40 tick target = -$7,715
80 tick target = -$10,655
120 tick target = -$6,915
16h00 exit = -$5,120

Results using just one contract with the following profit targets:
40 tick target = -$5,015
80 tick target = -$6,475
120 tick target = -$1,760
16h00 exit = -$2,250

All of the profit targets still yielded negative results, but averaging down yielded significantly worse results - taking stops on higher risk, when profits were not always realised on the bigger position just hurt returns significantly. At the same time, with the initial batch of "faulty" testing, pyramding as a trade works suffered from the same problem.

In view of these results, I think that trading multiple contracts is not as essential to trading success as I previously thought. Trading a fixed number of contracts could potentially lead to better results than adding to a position. Scaling out could smooth the equity curve, but the dip in profitability in 80 tick targets makes me unsure of how I would actually scale out.

@Tap In - Apologies if I side-tracked your journal a bit, but based on the previous discussions, I thought you would appreciate seeing these results.

Attached Files
Elite Membership required to download: CL Short-term 20t stop add.xlsx
Reply With Quote
The following 4 users say Thank You to grausch for this post:
  #808 (permalink)
 
ratfink's Avatar
 ratfink 
Birmingham UK
Market Wizard
 
Experience: Intermediate
Platform: NinjaTrader
Broker: TST/Rithmic
Trading: YM/Gold
Posts: 3,633 since Dec 2012
Thanks Given: 17,423
Thanks Received: 8,425


grausch View Post
In any case, I have redone the spreadsheets using adding in an easy way to test whether adding on retracements could work. Entries were determined by straddling price at 2am. I have not retested every single system I tried, since it takes quite a while and most of the results were not really that exciting. Trades were held until stops or targets were hit or until 16h00.

This is interesting but doesn't that sort of arbitrary entry condition make the whole exercise rather inconclusive? Unless I'm missing something.

Cheers

Travel Well
Visit my NexusFi Trade Journal Reply With Quote
The following 4 users say Thank You to ratfink for this post:
  #809 (permalink)
 grausch 
Luxembourg, Luxembourg
 
Experience: Advanced
Platform: TWS
Broker: Interactive Brokers
Trading: Stocks
Posts: 494 since May 2012
Thanks Given: 1,731
Thanks Received: 1,159


ratfink View Post
This is interesting but doesn't that sort of arbitrary entry condition make the whole exercise rather inconclusive? Unless I'm missing something.

Cheers

Well, the idea was to identify a breakout with the hope that it would lead to a larger trend. I had used Donchian channels in some tests, VWAP straddles once price was within a reasonable range, VWAP crossovers and price straddles at specific times. These lead to easier back testing than trend lines and support and resistance lines, which is why I chose them.

As stated before, I did not test any counter-trend strategies as I found all of the indicators very unreliable in the past. Perhaps counter-trend strategies would benefit more from averaging down. I do not know - I merely investigated one aspect. I do think that the result should at the very least should raise the question of whether scaling in is that much more superior to trading a fixed number of contracts.

Reply With Quote
The following 4 users say Thank You to grausch for this post:
  #810 (permalink)
 
ratfink's Avatar
 ratfink 
Birmingham UK
Market Wizard
 
Experience: Intermediate
Platform: NinjaTrader
Broker: TST/Rithmic
Trading: YM/Gold
Posts: 3,633 since Dec 2012
Thanks Given: 17,423
Thanks Received: 8,425



grausch View Post
As stated before, I did not test any counter-trend strategies as I found all of the indicators very unreliable in the past. Perhaps counter-trend strategies would benefit more from averaging down. I do not know - I merely investigated one aspect. I do think that the result should at the very least should raise the question of whether scaling in is that much more superior to trading a fixed number of contracts.

I would certainly expect those sort of standard indicators to perform as you found. I am also a scaling-in sceptic, but I have to be an agnostic until we can test it against a daily thesis, rather than arbitrary levels. And nope, as you said, that's not easy. Enough smart traders make it work to decide for themselves, but the rest of us still don't know if that's because scaling-in is inherently a good idea or because they just have good theses, although we do have our suspicions...

Cheers

Travel Well
Visit my NexusFi Trade Journal Reply With Quote
The following 4 users say Thank You to ratfink for this post:





Last Updated on December 28, 2019


© 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