Not setting a spread on a backtest.

• 01/23/2019 at 7:06 PM #89482
  Shitforbrains

  Hi,

   

  Sorry if this is a stupid question, but what would the difference be if I backtest a system with a 1 pip spread and 25 pip stop-loss vs. just setting a 26 pip stop-loss?

  Regards,

  

  Shitforbrains

  Participant

  Topics: 4

  Replies: 27

  Been thanked: 0 times
  01/23/2019 at 7:47 PM #89488
  GraHal

  1/2 / half / 0.5  a pip? 🙂

  You could try it in less than 5 mins?

  Let us know?

  I’ll try it when I next have my Platform open. I just closed all 4 Platforms as everything was getting jumpy due to RAM Hogging!

  

  GraHal

  Participant

  Topics: 204

  Replies: 8461

  Been thanked: 2057 times
  01/24/2019 at 9:13 AM #89531
  Shitforbrains

  Yep there’s a difference. 🙂 Have a look at these pics. And also the code:

  DefParam FlatBefore = 090000 DefParam FlatAfter = 130000 Defparam cumulateorders=false xClose = (Open+High+Low+Close)/4 if(barindex>2) then xOpen = (xOpen[1] + xClose[1])/2 endif c1 = xClose>xOpen AND xClose[1]<xOpen[1] c2 = xClose<xOpen AND xClose[1]>xOpen[1] if c1 AND Not LongOnMarket then BUY 1 SHARE AT MARKET EXITSHORT AT MARKET endif if c2 AND Not ShortOnMarket then SELLSHORT 1 SHARE AT MARKET SELL AT MARKET Endif SET STOP $LOSS 95 SET TARGET $PROFIT 85

  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

  DefParam FlatBefore = 090000   DefParam FlatAfter = 130000   Defparam cumulateorders=false     xClose = (Open+High+Low+Close)/4     if(barindex>2) then xOpen = (xOpen[1] + xClose[1])/2 endif   c1 = xClose>xOpen AND xClose[1]<xOpen[1] c2 = xClose<xOpen AND xClose[1]>xOpen[1]   if c1 AND Not LongOnMarket then BUY 1 SHARE AT MARKET EXITSHORT AT MARKET endif   if c2 AND Not ShortOnMarket then SELLSHORT 1 SHARE AT MARKET SELL AT MARKET Endif     SET STOP $LOSS 95   SET TARGET $PROFIT 85

  The first pic shows a backtest with 1 pip spread on the DAX. Stop loss 95, limit 85. The second pic shows the same system but with stop loss 96 and limit 84. I always get better results when I don’t define the spread. Any idea how there can be a difference between these two?

  

  Shitforbrains

  Participant

  Topics: 4

  Replies: 27

  Been thanked: 0 times
  01/24/2019 at 9:54 AM #89539
  jebus89

  Spread = cost.

  You need to pay for every trade(in and out) that you make.
  If you code it “into the stop loss” then ur not seeing that cost on every trade that does not hit stop loss.

  lets say you take a trade and it hits target. if you have your “spread” coded into your stop loss then that trade would be free, when in fact in real life that trade would have cost u the spread for in and out.

  1 user thanked author for this post.

  Shitforbrains

  

  jebus89

  Participant

  Topics: 118

  Replies: 586

  Been thanked: 238 times
  01/24/2019 at 10:59 AM #89552
  GraHal

  Shitforbrains wrote:

  The second pic shows the same system but with stop loss 96 and limit 84.

  Are you saying you had no spread entered in the box in the above example?

  If yes to above , then isn’t above equivalent to a spread of 2 (whereas you had spread = 1 when Stop loss 95, limit 85).

  Coffee has not woken my brain yet my suggestion above may be wrong?

  

  GraHal

  Participant

  Topics: 204

  Replies: 8461

  Been thanked: 2057 times
  01/24/2019 at 12:09 PM #89559
  Shitforbrains

  jebus89 wrote:

  Spread = cost.

  You need to pay for every trade(in and out) that you make.

  If you code it “into the stop loss” then ur not seeing that cost on every trade that does not hit stop loss.

  lets say you take a trade and it hits target. if you have your “spread” coded into your stop loss then that trade would be free, when in fact in real life that trade would have cost u the spread for in and out.

  Okey, great answer. So I could basically go through each and every trade that doesn’t hit stop or limit to calculate how much the spread cost would be for those specific trades.

  

  Shitforbrains

  Participant

  Topics: 4

  Replies: 27

  Been thanked: 0 times
  01/24/2019 at 12:26 PM #89561
  jebus89

  Easiest way if ur unable to add spread/cost into your backtest then you can basicly just take your amount of trades (correct me if im wrong here anyone) and mulitply them with the average spread, then multiply that with 2.

  100 trades * 1 pip (average spread) = 100 pips * 2 (1 for entry and 1 for exit) => 200 pips cost for doing 100 trades

  Note that for example IG have different average spread depending on the time of the day.

  Guessing this method is not the most accurate one because the spread does ultimately change your target and stop loss, depending on how youve coded ur system.

  if ur using PRT just put the spread in.

  

  jebus89

  Participant

  Topics: 118

  Replies: 586

  Been thanked: 238 times
  01/24/2019 at 12:33 PM #89562
  Shitforbrains

  GraHal wrote:

  Shitforbrains wrote:

  The second pic shows the same system but with stop loss 96 and limit 84.

  Are you saying you had no spread entered in the box in the above example?

  If yes to above , then isn’t above equivalent to a spread of 2 (whereas you had spread = 1 when Stop loss 95, limit 85).

  Coffee has not woken my brain yet my suggestion above may be wrong?

  Hmm, but both system would stop out at the same level if I run the backtests like this right? The one that has 86 pip stop loss would stop out on the same price as with 85 pip stop loss with spread included.

  

  Shitforbrains

  Participant

  Topics: 4

  Replies: 27

  Been thanked: 0 times
  01/24/2019 at 1:48 PM #89573
  jebus89

  You need to pay for entry as well as exit. if your stopping urself out 1 pip less, or taking a profit target 1 pip less, then thats just 1 pip. You would need 2. But this is affecting your backtest because your now moving your stop loss and target to try to calculate the spread, when you could actually just add the spread into backtest.. Also if you have trades that do not hit stop or target (but exit via an actual exit or trailing stop loss) then ur calculations are off again.

   

  just add the spread and accept the hardship hehe

   

  Example: lets say your “range” is profit target – stop loss = 50 pips. if you now start to remove pips to calculate for the spread ur new range would be 46, that might be a very big change in your system depending on ur pip pr trade on avg. (because ur removing 2 pips from the target, but also from your stop loss, leaving u with a much smaller range than your system normally would have if you just add the spread in the backtest.

  

  jebus89

  Participant

  Topics: 118

  Replies: 586

  Been thanked: 238 times
  01/24/2019 at 1:59 PM #89575
  GraHal

  jebus89 wrote:

  100 trades * 1 pip (average spread) = 100 pips * 2 (1 for entry and 1 for exit) => 200 pips cost for doing 100 trades

  Unless my coffee STILL hasn’t worked through 🙂 if the spread = 1 then that is all you pay re spread costs?

  Ask – Bid = 1 so you buy at the Ask and sell at the Bid, cost 1 pip?

  

  GraHal

  Participant

  Topics: 204

  Replies: 8461

  Been thanked: 2057 times
  01/25/2019 at 11:46 AM #89633
  Shitforbrains

  All in all I think it’s best not to mess around with the spread considering there will be a lot of trades closed without hitting stops or limits.
  Thank you guys for the answers. 🙂

  

  Shitforbrains

  Participant

  Topics: 4

  Replies: 27

  Been thanked: 0 times
• Author
  Posts