investor behavior

• 01/12/2018 at 7:10 PM #59247
  JohnScher

  I am concerned with the question of how investor behaviour behaves.

  In a simple code below, I have found with the 4H timebase that Monday morning investor behavior is different than Tuesday morning, compare the attachments.

  If I set the time base to 4H, the result is much better.

  I have the assumption that on certain weekdays and weekdays people have different investor behaviour at certain times.

  I’d like to use that in trading.
  So I wanted to ask whether someone here, or someone else, has an indicator or is able to program an indicator that, similar to the seasonal patterns from the Pathfinder system, reflects the timebased patterns during the week in investor behavior.

   

  P. S. The main code is the “negative”, i. e. the short version of the “DailyOpenLong”, which lives from the RepulseStop in a way that is not generally recognized.

   

  First : Storm with timebase, german time, gmt +1, Dax 1 Euro 4H

  Second (next post): Storm without timebase, german time, gmt +1, Dax 1 Euro 4H

   

  Please check it yourself.

   

   

   

  4H Short with RepulseStop timebased

  //------------------------------------------------------------------------- // Maincodce STORM ... ShorT On Rising Markets // 4H Short Repulse Timebased //------------------------------------------------------------------------- //------------------------------------------------------------------------- // Haupttode : 4H Short Repulse //------------------------------------------------------------------------- // Dax 1 Euro Mini // TimeFrame 4H // maincode: 4H Short with Repulse-Stop // created by JohnScher // defparam flatafter = 210000 defparam cumulateorders = true Once position = 1 c1 = cci [21] (close) >-95 // to get out some special situations c2 = TR (close) >24 // nice condition c3 = repulse [3] (close) <0.1 // mine t1 = dayofweek = 1 and time = 130000 // timebase t2 = dayofweek = 1 and time = 170000 // timebase t3 = dayofweek = 2 and time = 090000 // timebase t4 = dayofweek = 2 and time = 130000 // timebase t5 = dayofweek = 3 and time = 130000 // timebase t6 = dayofweek = 4 and time = 090000 // timebase t7 = dayofweek = 4 and time = 130000 // timebase t8 = dayofweek = 5 and time = 090000 // timebase TradingTimeShort = t1 or t2 or t3 or t4 or t5 or t6 or t7 or t8 IF TradingTimeShort Then IF c1 and c2 and c3 Then IF Exponentialaverage [1] (close) < ExponentialAverage [4] (close) THEN sellshort position CONTRACTS AT MARKET Endif Endif Endif If repulse [3] (close) >0.1 then Exitshort at market Endif // not tested // on other major indizes // with Seasonalpatternmultipler from Pathfinder-Systems // with some Re-invest-Stategies // greets JohnScher

  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57

  //------------------------------------------------------------------------- // Maincodce STORM ... ShorT On Rising Markets // 4H Short Repulse Timebased //------------------------------------------------------------------------- //------------------------------------------------------------------------- // Haupttode : 4H Short Repulse //------------------------------------------------------------------------- // Dax 1 Euro Mini // TimeFrame 4H // maincode: 4H Short with Repulse-Stop // created by JohnScher   // defparam flatafter = 210000 defparam cumulateorders = true       Once position = 1   c1 = cci [21] (close) >-95 // to get out some special situations c2 = TR (close) >24 // nice condition c3 = repulse [3] (close) <0.1 // mine       t1 =  dayofweek = 1 and time = 130000 // timebase t2 =  dayofweek = 1 and time = 170000 // timebase t3 =  dayofweek = 2 and time = 090000 // timebase t4 =  dayofweek = 2 and time = 130000 // timebase t5 =  dayofweek = 3 and time = 130000 // timebase t6 =  dayofweek = 4 and time = 090000 // timebase t7 =  dayofweek = 4 and time = 130000 // timebase t8 =  dayofweek = 5 and time = 090000 // timebase   TradingTimeShort = t1  or t2 or t3 or t4 or t5 or t6 or t7 or t8     IF TradingTimeShort Then IF c1 and c2 and c3  Then IF Exponentialaverage [1] (close) < ExponentialAverage [4] (close) THEN sellshort position CONTRACTS AT MARKET Endif Endif Endif     If repulse [3] (close) >0.1 then Exitshort at market Endif   // not tested // on other major indizes // with Seasonalpatternmultipler from Pathfinder-Systems // with some Re-invest-Stategies     // greets JohnScher

   

   

   

  Translated with http://www.DeepL.com/Translator

   

   

  

  JohnScher

  Participant

  Topics: 44

  Replies: 336

  Been thanked: 84 times
  01/12/2018 at 7:18 PM #59251
  JohnScher

  And here without timebase

   

  4H Short RepulseStop without timebased

  //------------------------------------------------------------------------- // Maincodce STORM ... ShorT On Rising Markets // 4H Short Repulse Timebased //------------------------------------------------------------------------- //------------------------------------------------------------------------- // Haupttode : 4H Short Repulse //------------------------------------------------------------------------- // Dax 1 Euro Mini // TimeFrame 4H // Haupttode : 4H Short with Repulse-Stop // treated by JohnScher // tested 08.04.2013 - 26.11.17 with PRT 10.3 // 2 Points Spread // won 161 loose 212 // max DrawDown 853 Euro max RunUp 5.035 // winnings total 4,473 Euro defparam flatafter = 210000 defparam cumulateorders = true Once position = 1 c1 = cci [21] (close) >-95 c2 = TR (close) >24 c3 = repulse [3] (close) <0.1 TradingDayShort = dayofweek = 1 or dayofweek = 2 or dayofweek = 3 or dayofweek = 4 or dayofweek = 5 TradingTimeShort = time = 090000 or time = 130000 or time = 170000 If TradingDayShort and TradingTimeShort Then IF c1 and c2 and c3 Then IF Exponentialaverage [1] (close) < ExponentialAverage [4] (close) THEN sellshort position CONTRACTS AT MARKET Endif Endif Endif If repulse [3] (close) >0.1 then Exitshort at market Endif // not tested // on other major indizes // with Saisonalpatternmultipler from Pathfinder-Systems // with some Re-invest-Stategies // greets JohnScher

  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54

  //------------------------------------------------------------------------- // Maincodce STORM ... ShorT On Rising Markets // 4H Short Repulse Timebased //------------------------------------------------------------------------- //------------------------------------------------------------------------- // Haupttode : 4H Short Repulse //------------------------------------------------------------------------- // Dax 1 Euro Mini // TimeFrame 4H // Haupttode : 4H Short with Repulse-Stop // treated by JohnScher   // tested 08.04.2013 - 26.11.17 with PRT 10.3 // 2 Points Spread // won 161 loose 212 // max DrawDown 853 Euro max RunUp 5.035 // winnings total 4,473 Euro     defparam flatafter = 210000 defparam cumulateorders = true       Once position = 1   c1 = cci [21] (close) >-95 c2 = TR (close) >24 c3 = repulse [3] (close) <0.1       TradingDayShort = dayofweek = 1 or dayofweek = 2 or dayofweek = 3 or dayofweek = 4 or dayofweek = 5 TradingTimeShort = time = 090000 or time = 130000 or time = 170000 If TradingDayShort and TradingTimeShort Then IF c1 and c2 and c3  Then IF Exponentialaverage [1] (close) < ExponentialAverage [4] (close) THEN sellshort position CONTRACTS AT MARKET Endif Endif Endif     If repulse [3] (close) >0.1 then Exitshort at market Endif   // not tested // on other major indizes // with Saisonalpatternmultipler from Pathfinder-Systems // with some Re-invest-Stategies   // greets JohnScher

   

  

  JohnScher

  Participant

  Topics: 44

  Replies: 336

  Been thanked: 84 times
  01/12/2018 at 8:03 PM #59273
  GraHal

  Weren’t the … seasonal patterns from the Pathfinder system taken … taken from a website? I cant recall the website name now.

  You can put Times in the Optimiser you know?

  GraHal

  PS Here’s a bit on Time of the Day re EurUSD

  https://www.dukascopy.com/fxcomm/fx-article-contest/?Seasonal-Cycles-In-Fx-Trading=&action=read&id=2018

  

  GraHal

  Participant

  Topics: 203

  Replies: 8374

  Been thanked: 2024 times
  01/12/2018 at 8:07 PM #59274
  mr blue

  Investors don’t like Monday’s – always keep in mind. So you’re on the right path my friend.

   

  

  mr blue

  Participant

  Topics: 10

  Replies: 77

  Been thanked: 14 times
  01/13/2018 at 12:22 PM #59333
  Vonasi

  Some days of the week are busier than others, some days of the month are busier than others and some hours are busier than others. If you put yourself in the mind of a professional trader then you will see repetitive behaviour.

  Monday – ‘Think I’ll just sit back a bit and see what the markets are doing before getting in’.

  Friday – Think I’ll just close down my positions for the weekend just in case’

  Last Day of Month – ‘Think I’ll cash in so my account looks good and I’ll get a big bonus’

  Market Opening – ‘Think I’ll just do a few stop raids and see what I can pick up before lunch’

  Market Closing – I don’t want to hold some of this riskier stuff overnight while I’m sleeping so I’ll sell up and get back in tomorrow’.

  Lunchtime – ‘Forget the trading I’m hungry so I’m off to the bistro’

  Asian Hours – ‘All the big players are asleep’

  and so on……

  Whether it helps to make profits I’m not sure. I think probability and risk management have more chance on their own – but then there is no point trading a market that isn’t moving because everyone is at lunch I guess.

  3 users thanked author for this post.

  Inertia, GraHal, mr blue

  

  Vonasi

  Moderator

  Topics: 192

  Replies: 5088

  Been thanked: 1810 times
  01/13/2018 at 12:29 PM #59335
  mr blue

  LOL Vonasi – perfect summarized !

   

  

  mr blue

  Participant

  Topics: 10

  Replies: 77

  Been thanked: 14 times
  02/04/2018 at 2:48 PM #61431
  JohnScher

  Hello again.

  I wasn’t really interested in the behaviour of the BigBoys and their boredom.

  I was more concerned with the fact that investor behaviour on a Monday morning is different than on a Tuesday noon.

  Did you notice something like that too?

   

  What I mean is something like this….

  Code 1 is intended to effect Monday 090000 Long to Monday 130000.

   

  Long Montag 090000

  IF OpenDayofweek = 1 and Time = 090000 then buy 1 contract at market Endif If Time = 130000 Then sell at market Endif

  1 2 3 4 5 6 7

  IF OpenDayofweek = 1 and Time = 090000 then buy 1 contract at market Endif   If Time = 130000 Then sell at market Endif

  Result see attachment

   

  Code 2 is intended to effect Tuesday 130000 Long to Tuesday 170000

  Long Dienstag 130000

  IF OpenDayofweek = 2 and Time = 130000 then buy 1 contract at market Endif If Time = 170000 Then sell at market Endif

  1 2 3 4 5 6 7

  IF OpenDayofweek = 2 and Time = 130000 then buy 1 contract at market Endif   If Time = 170000 Then sell at market Endif

  Result see attachment

   

  Amazing or?

   

   

   

  

  JohnScher

  Participant

  Topics: 44

  Replies: 336

  Been thanked: 84 times
  02/04/2018 at 4:44 PM #61443
  Vonasi

  Unfortunately the future does not always pay any attention to the past.

  You could easily convert your little test code above to analyse each market just by making the day and time optimized variables and then run a backtest. This would then tell you which have been the most profitable hours and in which days in your sample period. The sample may be too small to be conclusive. The only way to prove if this information has any value at all is to then forward test by running a strategy that just buys or sells in those hours and days and then you can see if the future is willing to copy the past. If it does then this could be used to maybe add an edge to a strategy by only letting it trade the most likely hours on the most likely days to be profitable.

  Forward testing is boring as we would all like to get rich today based on what happened yesterday but it is a necessary evil that we must learn to embrace!

  

  Vonasi

  Moderator

  Topics: 192

  Replies: 5088

  Been thanked: 1810 times
  02/04/2018 at 5:37 PM #61449
  JohnScher

  Na hi.

  “This would then tell you which have been the most profitable hours and in which days in your sample period.”

  …

  That is the intention.

   

   

  “The only way to prove if this information has any value at all is to then forward test by running a strategy that just buys or sells in those hours and days and then you can see if the future is willing to copy the past…”

  Yes, all our plans here are based on the same principle.
  We test what would have been in the past if our program had run
  and then we take the program and run it in demo or live as forward testing

  but

  in the moment we complete the forward test, we were just testing the past.

   

  Regardless of this, I would like to know who can program such an indicator, which shows (x-units) on the basis of past data, has the price risen or dropped between 090000 and 130000 on Monday, if so, what percentage? How did the course from Monday 130000 to Monday 170000, etc. behave?
  Of course, one can then generalize such an indicator and base it on each value and each time unit.

  That´s it what i m looking for.

   

  Kind

   

   

   

  

  JohnScher

  Participant

  Topics: 44

  Replies: 336

  Been thanked: 84 times
  02/04/2018 at 5:51 PM #61452
  Vonasi

  Not an indicator but the strategy I suggested is attached as an ITF file. Run it and the Optimize Report will show which have been the best days/hours for long positions. It should be easily adapted for short positions if you wish.

  IF OpenDayofweek = D and Time = T then buy 1 contract at market Endif If OnMarket Then sell at market Endif

  1 2 3 4 5 6 7

  IF OpenDayofweek = D and Time = T then buy 1 contract at market Endif   If OnMarket Then sell at market Endif

   

  1 user thanked author for this post.

  Inertia

  

  Vonasi

  Moderator

  Topics: 192

  Replies: 5088

  Been thanked: 1810 times
  02/04/2018 at 6:44 PM #61471
  Vonasi

  Short version. ITF attached.

  It would probably make sense to analyse one day at a time (by setting D to the required day) and then chart each days results in Excel to see if there is any sort of rhythm to the results. I think that the % of wins is of more interest than the actual amount won as it only takes one big bar in any particular hour of any particular day to skew the results. Any % of wins over 50% is of interest as 50% can be achieved by a monkey with a pin!

  IF OpenDayofweek = D and Time = T then SellShort 1 contract at market Endif If OnMarket Then ExitShort at market Endif

  1 2 3 4 5 6 7

  IF OpenDayofweek = D and Time = T then SellShort 1 contract at market Endif   If OnMarket Then ExitShort at market Endif

   

  

  Vonasi

  Moderator

  Topics: 192

  Replies: 5088

  Been thanked: 1810 times
  02/04/2018 at 8:25 PM #61477
  GraHal

  Thanks Vonasi

  Sorry JohnScher I couldn’t resist! 🙂

  Seems most wins are close to 50% so no pattern maybe??  Or do we need to analyse more?

  What you reckon Vonasi??

  Cheers
  GraHal

  

  GraHal

  Participant

  Topics: 203

  Replies: 8374

  Been thanked: 2024 times
  02/04/2018 at 8:47 PM #61485
  GraHal

  If anything, 21:00 is a good time to short the eurusd but that is probably because a lot of traders exit at 21:00 (GMT / UTC +0) so they don’t hold overnight?

  

  GraHal

  Participant

  Topics: 203

  Replies: 8374

  Been thanked: 2024 times
  02/04/2018 at 10:48 PM #61497
  Vonasi

  I think that Forex is a probably more of a problem when searching for this sort of investor behaviour as it is more of a 50/50 casino IMHO. Indices I feel would be more likely to have some sort of investor behaviour time based bias. Having said that my attempts at using Excel on the DAX to try to find some sort of pattern have just resulted in an over-whelming feeling of randomness. I fear the data sample of 100k bars is too small to be of any use and even if I tested on more I fear the result would be similar.

  I used OpenTime rather than Time to make sorting the order more logical and highlighted the top 25% best results. Chart and Excel book attached.

  1 user thanked author for this post.

  GraHal

  

  Vonasi

  Moderator

  Topics: 192

  Replies: 5088

  Been thanked: 1810 times
  02/05/2018 at 1:04 AM #61506
  verdi55

  Trade DAX on short term time scales from Monday through Thursday between 9 and 10 am. And on Friday afternoon, after market opening in US. This is pretty regular.

  

  verdi55

  Participant

  Topics: 2

  Replies: 311

  Been thanked: 84 times
• Author
  Posts