NexusFi: Find Your Edge


Home Menu

 





The two sides of probability


Discussion in Traders Hideout

Updated
      Top Posters
    1. looks_one Massive l with 4 posts (24 thanks)
    2. looks_two Quick Summary with 1 posts (0 thanks)
    3. looks_3 bobwest with 1 posts (2 thanks)
    4. looks_4 Trailer Guy with 1 posts (4 thanks)
      Best Posters
    1. looks_one Tap In with 7 thanks per post
    2. looks_two Massive l with 6 thanks per post
    3. looks_3 iantg with 5 thanks per post
    4. looks_4 Trailer Guy with 4 thanks per post
    1. trending_up 3,529 views
    2. thumb_up 50 thanks given
    3. group 5 followers
    1. forum 11 posts
    2. attach_file 0 attachments




 
Search this Thread

The two sides of probability

  #1 (permalink)
 pema83 
Bellheim Germany
 
Experience: None
Platform: Sierra Chart
Trading: Emini
Posts: 11 since Jan 2017
Thanks Given: 9
Thanks Received: 5

Hello,

in the thread
a few members pointed out that even in a 50:50 chance market you can make money if you "bet" in favor of a higher reward. Meaning something in the line of target 2 points versus stop 1 point.

Someone posted nice plots of account developments having 40:60 or 30:70 win loss relations.

I have a question wich bugs me every time I read something in that direction.
If you have an automated system which purely on a random basis opens a long or a short position with a 1:2 risk:reward relation. Isn't any attempt to increase the monetary outcome by adjusting the risk:reward relation automatically impacting your win loss trade count in a negative way? With a risk reward ration of 3:1 the market has to walkt 3 times the "distance" in the right direction. One could also say the the chance of hitting your stop is 3 times as high as hitting your target. The same should apply to a not randomly opened position but also on a position which is for example long on an ascending moving average. Still the smaller stop room affects your chance of hitting target before stop in a bad way.

I'm not doubting the scientific correctness of "just increase your risk reward ratio" but where is the practical use if you reduce the occurrence of a positive outcome by increasing the reward.

I'm lacking a the statistical background, some vocabulary and englisch is not my main language. I hope the message comes through and I don't sound harsh.

Started this thread Reply With Quote
Thanked by:

Can you help answer these questions
from other members on NexusFi?
New Micros: Ultra 10-Year & Ultra T-Bond -- Live Now
Treasury Notes and Bonds
Are there any eval firms that allow you to sink to your …
Traders Hideout
Futures True Range Report
The Elite Circle
ZombieSqueeze
Platforms and Indicators
Build trailing stop for micro index(s)
Psychology and Money Management
 
  #3 (permalink)
 iantg 
charlotte nc
 
Experience: Advanced
Platform: My Own System
Broker: Optimus
Trading: Emini (ES, YM, NQ, ect.)
Posts: 408 since Jan 2015
Thanks Given: 90
Thanks Received: 1,147


There is no way to overcome the obvious points you mentioned:

Option 1. If your higher win rate comes at the expense of taking a larger risk, this will eventually just average out to 50/50. But you you will still pay all your transaction costs, spread costs etc.

Option 2. If you try to fudge the equation the opposite way you will have similar results. If you use tighter stops and larger profit targets you will have a lower win rate, but the value of the win rate will be higher. All things random and even, this will still average out to 50/50. You will still end up losing because of transaction costs, spread costs, etc.

In terms of how you exploit any of this and make money...... If you can correctly gauge the volatility, you can apply one or the other strategies at the right time and come out ahead.

1. In low volatility periods. You won't see huge moves, so if you set your stop outside of the range the market is moving it will rarely get hit. Option 1 will beat a low volatility market using random entries.

2. In high volatility periods, the market will have as much of a chance of moving 5 ticks as 10 ticks for example. So in these periods you go with option 2. Again, using random entries, all other factors even, you can mathematically be profitable doing this.


I had an old journal on here that describes this type of system in more detail. But basically you are betting on long term volatility patterns, not direction.

Best of luck.

In the analytical world there is no such thing as art, there is only the science you know and the science you don't know. Characterizing the science you don't know as "art" is a fools game.
Visit my NexusFi Trade Journal Reply With Quote
Thanked by:
  #4 (permalink)
 
Massive l's Avatar
 Massive l 
OR/USA
Legendary /NQ Trader
 
Experience: None
Posts: 2,129 since Mar 2011
Thanks Given: 1,859
Thanks Received: 5,106

I have a hard time believing any professional trader uses arbitrary r/r ratios in their trading. r/r is a byproduct of the system, not the foundation.

However, you are correct. When you change arbitrary r/r ratios you're also changing your winning %.

Visit my NexusFi Trade Journal Reply With Quote
Thanked by:
  #5 (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


pema83 View Post
Hello,

in the thread
a few members pointed out that even in a 50:50 chance market you can make money if you "bet" in favor of a higher reward. Meaning something in the line of target 2 points versus stop 1 point.

Someone posted nice plots of account developments having 40:60 or 30:70 win loss relations.

In the post you refer to, the math is correct but the assumptions are not realistic and therefore do not work in the real world.

We all know that the farther out the target the less often the target will be reached. In the real world it is actually worse than that.

In the real world, especially for an intraday trader with finite daily trading time windows, not only will a farther out target be reached less often, but the percentage of time it is reached will continually degrade well below the percentage required to breakeven as the target is moved farther out.

The farther out the target the less the chance of reaching it often enough to breakeven. Eventually a target will have a nearly zero percent chance of getting reached because it is larger than the largest daily range over a given time span.

For example, setting aside slippage, commissions and the nature of how a market order is filled compared to a limit order (a big ask, I know!), the percentages of wins required to breakeven in a set-it-and-forget-it scenario are:

Risk:reward
1:1 50%
1:2 33%
1:3 25%

This is simple math.

In the real world, based on empirical evidence that I've kept over hundreds of trades, if I am only 50% successful reaching 1R over a series of trades, my success rate in reaching the farther targets look more like the following:

1:1 50%
1:2 25%
1:3 13%

As you can see, the farther the target, the more degraded the success percentage. All scenarios beyond 1:1 are losers. Add in slippage, commissions and the nature of how a market order is filled compared to how a limit order is filled and it gets worse.

Reply With Quote
  #6 (permalink)
 Trailer Guy 
Aguanga, CA USA
 
Experience: Advanced
Platform: Ninja Trader 8
Broker: IB, NinjaTrader Brokerage, Schwab
Trading: ES
Posts: 215 since Sep 2014
Thanks Given: 68
Thanks Received: 387

1. The market is not a Gaussian probability cure, that is it does not have a bell shaped probability. It has a skewed curve with a fat tail. So as others have said your basic assumption is false.
2.How often a trade can happen is an issue on total returns. But if it is just a random system you need to look at what the average rotation is. Another way to look is ATR. Market ATR increases with the time for the bars. As others have said if you are inside of random movement and counting on a move outside of random you need luck.
3. In one of Van Tharp's books they used random entry with appropriate position sizing and stop loss. It made a little money. I have never seen an experiment with random entry and exit. What is that a double dart board system?
4. As others have said risk reward is only a part of the calculation. Going back to Van Tharp he has a very good suggestion. Compile the results (in R) for 100 trades and then do a histogram to see what your distribution looks like.
5. For us little people math is a tool, not a machine. It will help us craft a better trade but it will not guarantee it. Only Renaissance hedge fund has a math machine and they have never been able to build a second one.

Reply With Quote
Thanked by:
  #7 (permalink)
 
Massive l's Avatar
 Massive l 
OR/USA
Legendary /NQ Trader
 
Experience: None
Posts: 2,129 since Mar 2011
Thanks Given: 1,859
Thanks Received: 5,106

system not using arbitrary r/r

1.89 R/R 66% win $223 expectancy/contract/trade

0 - .85 14%
.86 - 1.70 17%
1.71 - 2.56 11%
2.57 - 3.42 11%
3.43 - 4.28 7.5%
4.29 - 5.13 4.3%
5.14 - 6.00 1.1%

Same system entries but with arbitrary R/R ratios

1:1 66% win $95 expectancy
2:1 50% win $98 expectancy
3:1 41% win $109 expectancy
4:1 36% win $141 expectancy
5:1 26% win $78 expectancy
6:1 23% win $90 expectancy

Visit my NexusFi Trade Journal Reply With Quote
  #8 (permalink)
 Grantx 
Reading UK
Legendary no drama Llama
 
Experience: None
Posts: 1,787 since Oct 2016
Thanks Given: 2,826
Thanks Received: 5,058


Trailer Guy View Post
1. The market is not a Gaussian probability cure, that is it does not have a bell shaped probability. It has a skewed curve with a fat tail.

The market doesnt move in a straight line and thats why you model it using linear regression. Thats where you will find your perfectly shaped probability curve.

Visit my NexusFi Trade Journal Reply With Quote
Thanked by:
  #9 (permalink)
 
bobwest's Avatar
 bobwest 
Western Florida
Site Moderator
 
Experience: Advanced
Platform: Sierra Chart
Trading: ES, YM
Frequency: Several times daily
Duration: Minutes
Posts: 8,162 since Jan 2013
Thanks Given: 57,343
Thanks Received: 26,267


Massive l View Post
I have a hard time believing any professional trader uses arbitrary r/r ratios in their trading. r/r is a byproduct of the system, not the foundation.

However, you are correct. When you change arbitrary r/r ratios you're also changing your winning %.


Massive l View Post
system not using arbitrary r/r

1.89 R/R 66% win $223 expectancy/contract/trade

0 - .85 14%
.86 - 1.70 17%
1.71 - 2.56 11%
2.57 - 3.42 11%
3.43 - 4.28 7.5%
4.29 - 5.13 4.3%
5.14 - 6.00 1.1%

Same system entries but with arbitrary R/R ratios

1:1 66% win $95 expectancy
2:1 50% win $98 expectancy
3:1 41% win $109 expectancy
4:1 36% win $141 expectancy
5:1 26% win $78 expectancy
6:1 23% win $90 expectancy

@Massive l, I'm trying to understand your numbers here, because I think they tell an important story. But can you just explain explicitly what you mean? I think it's that your results would have been cut short if you had killed the trades at pre-set R multiples, which I think is often true, but I don't really quite understand the numbers themselves. Can you say more, please?

Bob.

When one door closes, another opens.
-- Cervantes, Don Quixote
Reply With Quote
Thanked by:
  #10 (permalink)
 
Massive l's Avatar
 Massive l 
OR/USA
Legendary /NQ Trader
 
Experience: None
Posts: 2,129 since Mar 2011
Thanks Given: 1,859
Thanks Received: 5,106



bobwest View Post
@Massive l, I'm trying to understand your numbers here, because I think they tell an important story. But can you just explain explicitly what you mean? I think it's that your results would have been cut short if you had killed the trades at pre-set R multiples, which I think is often true, but I don't really quite understand the numbers themselves. Can you say more, please?

Bob.

I have non-arbitrary r/r results with the % of total wins making up each r/r range. I didn't feel like writing the code to give me exact 2:1, 3:1, etc. as the auto range that was spit out would suffice.

The bottom numbers show arbitrary r/r results.

The results show that using arbitrary r/r limits your profits and is not my preferred method of system building.
If I get a wild hair later I'll go with other way with r/r by keeping my profit the same while increasing risk.

I do use a max stop loss (only hit once in the last 100 trades and it's not that big), but $ or tick amount in profits changes based on that day's volume, bar height, time.

You can also see that 4:1 produces the highest expectancy ($92 less than non-arbitrary though).
However, it only wins 36% of the time. For my mental health, I prefer to win at least 50% of the time so even with the highest expectancy I still would never trade something that won 36% of the time. Maybe another +10% I might consider it.

Visit my NexusFi Trade Journal Reply With Quote




Last Updated on July 20, 2020


© 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