The Killing Floor | Page 6 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Oanda Killing CPU 3 replies

making a killing!! in this market 48 replies

making a killing!! in this market prt 2 71 replies

Off quotes killing my EA 0 replies

Spam filters are killing my chart alerts 0 replies

Attachments: The Killing Floor
Exit Attachments

The Killing Floor

  • #101
  • Jun 26, 2018 9:35am Jun 26, 2018 9:35am
Quoting Sis.yphus
{quote} You can drop all the well structured goodies you like here my friend Any well thought out and rationalized ideas/discussion are welcome as I want this thread to be one of the highest quality ones in terms of substance on this forum.
Yep...remember I am a neutral guy with no permanent home on the Factory. I am like living on "No Man's Land" until I reach "Land's End".

Speaking of Land's End, I have travelled many Lands End locations around the world. It is on my bucket list to visit each and every one of them (except the ones way up North and South...I am too old to freeze my b**ls LOL)
It is something so magical, refreshing and spiritual to visit such remote locations. I have already chosen my destination on where and how to depart :-)
Staying in my lane...
  • #102
  • Jun 26, 2018 11:29am Jun 26, 2018 11:29am
Choppy as day on EU today...

Heading out for a few hours.

Mood:
Inserted Video
  • #103
  • Jun 26, 2018 2:58pm Jun 26, 2018 2:58pm
Equity Progression

You ever wonder what your account growth would look like if you got your hands on a winning system? I bet you imagine it going straight up as it smoothly sails you into the lap of luxury... I am almost positive you don't imagine it would look how it really does.

Below is how your equity would progress using a really strong system over x amount of trades. This simulation uses a system that risks 5 units for 7 units with a 50/50 strike rate and a spread of 0.5 pips.

10 Trades:
Attached Image (click to enlarge)
Click to Enlarge Name: 10t.PNG Size: 27 KB


How many would keep trading after these 10 trades? Let's add 10 more...

20 Trades:
Attached Image (click to enlarge)
Click to Enlarge Name: 20t.PNG Size: 30 KB


How many would keep trading after these next 10 trades? How about 10 more...

30 Trades:
Attached Image (click to enlarge)
Click to Enlarge Name: 30t.PNG Size: 33 KB


How many would keep trading after these 10 additional trades? How about 10 more...

40 Trades:
Attached Image (click to enlarge)
Click to Enlarge Name: 40t.PNG Size: 34 KB


How many would keep trading after these 10 additional trades? "This thing has to be break even at best!" - Most would say...

How about 10 more?

50 Trades:
Attached Image (click to enlarge)
Click to Enlarge Name: 50t.PNG Size: 31 KB


Ah, now we are seeing something that is at least somewhat profitable looking. But it took 50 trades just to make a new equity high that was set in the first 10 trades... Does this system really have an edge? How about 50 more trades?

100 Trades:
Attached Image (click to enlarge)
Click to Enlarge Name: 100t.PNG Size: 42 KB


Now we are looking pretty good. This edge may be real! LUXURY HERE WE COME!

Let's do 100 more...

200 Trades:
Attached Image (click to enlarge)
Click to Enlarge Name: 200t.PNG Size: 43 KB


O man, I don't know anymore, this is looking a bit rocky. Is this edge even real? Let's do 100 more.

300 Trades:
Attached Image (click to enlarge)
Click to Enlarge Name: 300t.PNG Size: 47 KB


O yeah! Things are looking good again, let's do 100 more.

400 Trades:
Attached Image (click to enlarge)
Click to Enlarge Name: 400t.PNG Size: 48 KB


Yeah, baby! We are still climbing! How about 100 more?


500 Trades:
Attached Image (click to enlarge)
Click to Enlarge Name: 500t.PNG Size: 49 KB


Well, I think it is safe to say this edge is real...

Exercise:

Imagine each image is your equity curve progressing.

How many of you would have kept trading after those first 10 trades?

How about after the first 20?

How many of you would have tried to change or tweak something after the first 40 trades?

How many of you would be able to trust the system, and not change a single thing for the first 100, 200, 300+ trades...

The answer. Not many. Why is that?

I'd say its a mix between lack of conviction in the system and the lack of understanding of the central limit theorem and the law of large numbers... but that's just me.
  • #104
  • Jun 26, 2018 5:00pm Jun 26, 2018 5:00pm
Quoting Sis.yphus
Equity Progression Well, I think it is safe to say this edge is real... Exercise: Imagine each image is your equity curve progressing. How many of you would have kept trading after those first 10 trades? How about after the first 20? How many of you would have tried to change or tweak something after the first 40 trades? How many of you would be able to trust the system, and not change a single thing for the first 100, 200, 300+ trades... The answer. Not many. Why is that? I'd say its a mix between lack of conviction in the system and the lack...
Too few people would have.

Please indulge me.

I think the challenge is how many of those draw downs actually hit SL? You really have to know that with 1000 consecutive losses on a 1:1 leverage your equity can survive to fight another 500 trades of 36 pips SL.
Let it turn and find your entry.
  • #105
  • Jun 27, 2018 3:05am Jun 27, 2018 3:05am
How I tested Cutting losers short.
My entry is always buy on new H1 open candle. TP = 5p(Added 1p spread, so with spread 4p), SL = -35p(with spread -36p). Why entry H1 open and not MA's? Because it is simpler to code, and the same probability of entry: 50:50 , there is nothing magical about MA's.
So if I open trades with Tp and SL, and don't cut losers, wait for TP or SL, the stats are :
Trades:84600;
Wins:74828;
Losers:9772;
Win%: 88.45%;
Average pips per trade = -0.62p.
So after 10 minutes of candle open I got MAE stats(trades which lasts more than 10 minutes):
MAE winners: 5.1p;
MAE losers: 9.1p.
So, lets try to cut losers if within 10 minutes trade's MAE gets to >=20p(tried with 10p, 30p - result the same):
Trades:84600;
Wins:68723;
Losers:15877;
Win%: 81.23%;
Average pips per trade = -0.67p;
Average win pips: 4p;
Average loser pips: -20.86p.
As you can see, if you cut losers, you cut your winners too, and the result is the same, when you look at average pips per trade.
Of course, I could do it more complicated: look for MAE's after each minute, each 10 minutes. But some positive results should be seen from this simple analysis.
So maybe am I doing something wrong? Has someone else tried something similar?
  • #106
  • Jun 27, 2018 7:37am Jun 27, 2018 7:37am
Quoting IenDzi
How I tested Cutting losers short. My entry is always buy on new H1 open candle. TP = 5p(Added 1p spread, so with spread 4p), SL = -35p(with spread -36p). Why entry H1 open and not MA's? Because it is simpler to code, and the same probability of entry: 50:50 , there is nothing magical about MA's. So if I open trades with Tp and SL, and don't cut losers, wait for TP or SL, the stats are : Trades:84600; Wins:74828; Losers:9772; Win%: 88.45%; Average pips per trade = -0.62p. So after 10 minutes of candle open I got MAE stats(trades which lasts more...

You literally "test" the strategy using the wrong time frame, wrong opening conditions, scrapped all the tools altogether and their related cycles, and tried to 100% automate it despite me saying multiple times it wouldn't work...

Quoting IenDzi
So maybe am I doing something wrong?
And you are surprised it didn't work...

You MUST be trolling me right?
  • #107
  • Jun 27, 2018 8:21am Jun 27, 2018 8:21am
Quoting Sis.yphus
{quote}So if you don't like the R:R I am using, use whatever one you like that fits your style. I personally like to win a lot so I prefer the higher hit hates.
I didn't like the entry you were using, so I used my entry. Because, it's the same 50:50 result at the end, if I use MA's or open price.

Quoting Sis.yphus
{quote}Whatever r:r you use, just make sure you understand that the r:r does not provide you any advantage in terms of gaining an edge. All it does is change how you will manage your winners and losers, the average size of your winners and losers, and the distribution of your winners and losers.
I didn't want to test all your system. All I wanted to test, how much your edge increases managing your winners/losers in your written aproach. I think, it should be seen some micro edge no matter the entry you are using.
I am not trolling, sorry, if I understood wrong.
  • #108
  • Jun 27, 2018 9:15am Jun 27, 2018 9:15am
https://www.forexfactory.com/showthread.php?t=443430

This guy blew my mind some years back.

I just am interested in the psychology of trading like this, how do you get the sense of blood, how do you know or guess it's time to cut or time to stack?
Might be helpful if some more charts of actual trades were posted.
Let it turn and find your entry.
  • #109
  • Jun 27, 2018 9:23am Jun 27, 2018 9:23am
Quoting Intrepidpips
https://www.forexfactory.com/showthread.php?t=443430 This guy blew my mind some years back. I just am interested in the psychology of trading like this, how do you get the sense of blood, how do you know or guess it's time to cut or time to stack? Might be helpful if some more charts of actual trades were posted.
The r:r you choose does really depend on your psyche. For me, doing the reverse is much harder.

Combine the cycle data with the MAE/MFE data and you should have a really clear picture of which sections of the market you should be participating in.
  • #110
  • Jun 27, 2018 10:04am Jun 27, 2018 10:04am
Not bad aye?
Attached Image (click to enlarge)
Click to Enlarge Name: Screenshot1.png Size: 106 KB
  • #111
  • Jun 27, 2018 10:06am Jun 27, 2018 10:06am
Quoting Ivanyo
Not bad aye? {image}
Not bad at all my friend
  • #112
  • Edited 1:49pm Jun 27, 2018 10:50am | Edited 1:49pm
Probability of Profit in a Series
*New trader gets a hold of a profitable system. Takes 500 trades over a few weeks or months. Still loses*
Attached Image


In this last post, we looked at how our equity curve is likely to progress using a pretty good system. That gave us a good idea of what to expect, but it still leaves the question, "how many trades do I need to take in order to expect to be profitable?"

And that is a fine question to ask. The answer to that question will give you a lot of insight in terms of what conditions need to be present for you to consistently make money within a fixed period of time. As I have mentioned in other posts, this is how large trading desks like Goldman Sachs will make money 9 days out of 10, for years. This will also give a good way to benchmark your system and make sure it is actually performing how it should be.

Let's get into some numbers. I wrote a script that will do 100,000 simulations of a trading series lasting x amount of trades. It then looks at each of them to determine the probability of being in the money at the end of that series. This uses the same system in the last post. 50% Win Rate, 5 Pip SL, 7 Pip TP, and 0.5 Pip Spread.

WIN RATE: 50% | SL: 5 | TP: 7 | SPREAD: 0.5
Attached Image


Pretty amazing stuff. In any series of 10 trades, you only have a 61% chance of making money. In any series of 100 trades, you only have an 81% chance of making money. That means it would be totally normal to have a series of 100 trades where you end up losing money as a whole, out of every 500 trades you take. If 5 people pick this system up and each does 100 trades. One of them will lose money, and that will be perfectly normal from a statistical standpoint.

In order to be 95% confident you will make money in a specific amount of time, you need to be able to have 400 independent occurrences. Trading desks have so many traders so their net number of occurrences across the desk can easily be well over 500 trades. Same fore HFT algos scalping nano ticks. They don't make money every day because they cheat. They make money every day because they understand numbers.

WIN RATE: 93.5% | SL: 36 | TP: 4 | SPREAD: 0.5
Attached Image

Notice how the number differ from the other system, but not drastically or in some huge way. In some series lengths, this system is more likely to turn a profit, and in other series, the other system is more likely to see a profit. Both systems need 400 independent occurrences to have 95% confidence that you will see a profit at the end of the series.

Think about this: If we had 100 traders pick up the first system and 100 traders pick up the second system and each trade 400 trades. 5 traders from each group will still lose money after 400 trades with the winning systems, and that will be statistically normal.
  • #113
  • Jun 27, 2018 10:18pm Jun 27, 2018 10:18pm
Quoting Sis.yphus
Probability of Profit in a Series *New trader gets a hold of a profitable system. Takes 500 trades over a few weeks or months. Still loses* {image} In this last post, we looked at how our equity curve is likely to progress using a pretty good system. That gave us a good idea of what to expect, but it still leaves the question, "how many trades do I need to take in order to expect to be profitable?" And that is a fine question to ask. The answer to that question will give you a lot of insight in terms of what conditions need to be present for you...
Hi Sis.yphus,

Thanks for putting this information together - it does give interesting insights into the probability of profitable trades given the various scenarios. I must say, my initial trading session has yielded great results as you already know, so I am looking to refine my routine a little bit (what times to trade, how many trades per day etc) to put together a complete trading system that suits my lifestyle.

One question I had was whether you have the raw data that you used to put together the graphs on post #1. The data on the graphs look similar across all entry scenarios, and is hard to differentiate between the various entry methods. I would like to know how much better (or worse) the various methods are based on their entry criteria (so for instance, I am seeking data like "entering when price has closed below 200,50 and 10 MA and sloping down has a 0.5% better chance than when price has closed below 200 and 50 MA to result on a profitable trade").

Cheers
  • #114
  • Jun 28, 2018 12:00am Jun 28, 2018 12:00am
Quoting Sis.yphus
Probability of Profit in a Series *New trader gets a hold of a profitable system. Takes 500 trades over a few weeks or months. Still loses* {image} In this last post, we looked at how our equity curve is likely to progress using a pretty good system. That gave us a good idea of what to expect, but it still leaves the question, "how many trades do I need to take in order to expect to be profitable?" And that is a fine question to ask. The answer to that question will give you a lot of insight in terms of what conditions need to be present for you...
You are a rock star man. I have had so many discussions and arguments over this topic when folks would just advocate "trade higher timeframe", "have patience", "market is going to be around tomorrow", "be picky about trades" etc etc. Folks just don't get trading is all about numbers and not about reading price setups or analyzing entries or waiting for days/weeks for the right entry and big RR. I have used a similar approach with my automation as well..but without using hard stops for obvious reasons. I just rely on (time) frequency of profit taking to exceed the frequency of loss taking via active and decisive management of my own "order book" i.e available margin just like market makers are required to actively manage their book size. Same concept of large numbers are 'necessary' for casinos and bucket shops to maintain their contineous flow of tiny profits. There is a reason why there are 10000's of slot machines. There is a reason why retail brokers need 1000s of retail tiny accounts :-
Staying in my lane...
  • #115
  • Jun 28, 2018 7:41am Jun 28, 2018 7:41am
Today's trades - going to call it a day and relax.
Attached Image (click to enlarge)
Click to Enlarge Name: Screenshot1.png Size: 93 KB
  • #116
  • Jun 28, 2018 8:30am Jun 28, 2018 8:30am
Quoting Ivanyo
Today's trades - going to call it a day and relax. {image}
Very nice! Bagged your pips and now I see you playing some PUBG. I can dig it...

Quoting VEEFX
{quote} You are a rock star man. I have had so many discussions and arguments over this topic when folks would just advocate "trade higher timeframe", "have patience", "market is going to be around tomorrow", "be picky about trades" etc etc. Folks just don't get trading is all about numbers and not about reading price setups or analyzing entries or waiting for days/weeks for the right entry and big RR. I have used a similar approach with my automation as well..but without using hard stops for obvious reasons. I just rely on (time) frequency of profit...
Exactly sir!

Glad you find it all useful!

Understanding how the numbers play out on a large scale is a major key to building conviction in what you are doing.
  • #117
  • Jun 28, 2018 9:09am Jun 28, 2018 9:09am
Quoting Sis.yphus
{quote} Very nice! Bagged your pips and now I see you playing some PUBG. I can dig it... {quote} Exactly sir! Glad you find it all useful! Understanding how the numbers play out on a large scale is a major key to building conviction in what you are doing.
Haha yes...nice to play a few games to take the give the mind a little break!
  • #118
  • Jun 28, 2018 10:21am Jun 28, 2018 10:21am
a small contribution, maybe it will be useful to someone

Attached File(s)
File Type: pdf The Killing Floor - polish.pdf   3.4 MB | 1,112 downloads
  • #119
  • Jun 28, 2018 10:33am Jun 28, 2018 10:33am
Quoting BlackNapkins
a small contribution, maybe it will be useful to someone {image}
Thanks for translating those posts! I will link this on the front page.
  • #120
  • Jun 28, 2018 10:43am Jun 28, 2018 10:43am
Quoting Ivanyo
Today's trades - going to call it a day and relax. {image}
Can I ask a (no doubt daft) question?

Why are you having to hold trades so long? For instance, nearly two hours to make +1.1 pips on UJ.

I am honestly not trying to be cheeky or wind you up but when I run this (and other scalping of the same ilk), I am in and out in between one and five seconds on a good day.

Today, while sitting waiting for an eye exam appointment, I took four trades on my phone:

GU 1' 26" (One minute Twenty-Six seconds)
GU 9" (Nine seconds)
GU 25" (Twenty-Five seconds)
DAX 22" (Twenty-Two seconds)

If I am missing something blatantly obvious or stepping out of line, feel free to slap me about the head.
0 traders viewing now