No Trades when Spread > 'Normal'?

Forums › ProRealTime English forum › ProOrder support › No Trades when Spread > 'Normal'?

Viewing 15 posts - 31 through 45 (of 50 total)
1 2 3 4
  • 03/18/2020 at 6:33 PM #122492
    GraHal

    Lies, damned lies and statistics … I was fitting the figures to a scenario! 🙂

    I was thinking I was looking at real spread, but I am looking at the spread I had set in the backtester!

    So the middle line = 10 (9.96) in above images is due to me setting spread = 10 in the backtester!

    I was thinking slippage wouldn’t be the same on every trade and then a light came on! 🙂

    1 user thanked author for this post.
    avatar Paul
    03/18/2020 at 9:04 PM #122501
    Paul

    what about the following. Rough quick idea. A lot of potential problems, but post it anyway.

    Have a system running live in demo & keep the same system in de live demo making fake trades too.

     

    Then realtime in live demo compare them (not results but entry- and exitprice )  and if the (average?) difference increases about a preset number (caused by spread&slippage), stop trading. To make it easy it should run only in the lowest spread timezone for starters.

     

     

    03/19/2020 at 10:43 AM #122548
    GraHal

    Main problem  … we would need to be at the screen watching … if we do that then we can see what the spread is anyway in the top right hand corner.

    Or am I missing something??

     

    03/19/2020 at 1:53 PM #122576
    Paul

    Well, since spread can’t be retrieved automatically or you have to be behind the screen like you said, a workaround is to see spread & slippage (cause by volatility!) as a whole. It can’t work for the exit, but only for the entry’s.

    You can only take action after you are experiencing it, so, say you allow up 10 points worse entry-price on the dow (lowest spread times). Trades are executed live and the real entry-price is compared to the theoretical simulated entry-price in the same strategy.

    The theoretical price can be the open of the bar of the market order +/- know published IG spread.

    If the difference is bigger then i.e.the threshold of 10 points, spread & slippage effects the stoploss & the exit-price so much the strategy is then automatically stopped. (no restart)

    It would be better not to use the last trade, but the last i.e. 3-5 trades average differences.

    It’s like a plug & play code, just pop it in a strategy and set the threshold to stop trading.

    If a system works 24h with different spreads it’s much more difficult.

    Here’s an article to read about fake/simulated trading btw.

    https://www.prorealcode.com/blog/learning/how-to-improve-a-strategy-with-simulated-trades-1/

    03/19/2020 at 2:20 PM #122582
    GraHal
    Paul wrote:

    If a system works 24h with different spreads it’s much more difficult.

    Sounds very good Paul!

    I now understand, given the simulated trades within the overall code for the Strategy.

    Re the 24 hour difficulty / mixed spread values … we can just accept that when spread is > user set x points … the strategy doesn’t trade.

    03/19/2020 at 3:17 PM #122586
    Paul

    yes indeed and high slippage is also a big part to stop trading then.

    Here’s an early indicator for the dax I made which shows the spread. It captures the 24h problem.

    The question is , is it worth the effort? I think not, because it creates again a variable, when to stop trading. Once stopped, it can’t be restarted.  And it isn’t clear what’s the reason to stop, the high slippage or the increased spread or a mixture of both.

    Still nice to think about it though.

     

    03/19/2020 at 4:18 PM #122591
    GraHal
    Paul wrote:

    Once stopped, it can’t be restarted.

    Ah I need to re-read then as I thought the simulated trades would be coded into the Strategy and would stop trades (but not stop the Strategy) for x bars or until the next day at 08:00 or whatever would be set by the user?

    03/19/2020 at 4:29 PM #122595
    Paul

    mmm…..In doubt now 🙂

    You’re right, in the link I sent it could be restarted. But here compared is real vs simulated entryprice.

    My thought is there’s nothing to compare against to restart trading, because no live data is coming in anymore.

    It’s like when the strategy is stopped, then you go to the screen!

     

    03/19/2020 at 4:38 PM #122597
    fifi743

    The simplest is that PRT adds the variables SPREAD and the minimum distance of the stop

    03/19/2020 at 6:51 PM #122612
    GraHal
    Paul wrote:

    My thought is there’s nothing to compare against to restart trading, because no live data is coming in anymore.

    So just have it that it restarts anyway at a time (market open + 15 mins) when spread is historically at its lowest for the respective instrument and then the cycle starts over until it stops next time then rinse and repeat?

    I am still thinking the simulated trades are coded into the Strategy and that trading stops but the System does not stop … is that correct?

    03/19/2020 at 6:54 PM #122613
    GraHal
    fifi743 wrote:

    simplest is that PRT adds the variables SPREAD and the minimum distance of the stop

    These are controlled by IG (not PRT) and IG would not want us to have visibility as using spread and min distance is how IG balance their books.

    1 user thanked author for this post.
    avatar fifi743
    03/21/2020 at 1:32 PM #122823
    mr blue

    I was also looking for the spread information, but then I did a workaround using the averagetruerange …. if the volatiliy is extremly high, you can expect that the broker will increase the spread. also you can consider the trading hours. spreads are smaller during offical cash market hours.

    you find information about spreads and cash market hours on the ig homepage.

    hope that gives a littel help

     

    1 user thanked author for this post.
    avatar GraHal
    03/21/2020 at 2:13 PM #122826
    Paul
    GraHal wrote:

    I am still thinking the simulated trades are coded into the Strategy and that trading stops but the System does not stop … is that correct?

    Yes I think so. The thought is the simulated trades continue and the real trading stops while the system keeps running. There can be a switch that it restarts real trading next day and the comparison starts again.

    I’ve added fifi’s code I posted earlier in the renko strategy. It captures the volatility a bit. Averaging the difference between open and close. It needs to be smaller then xx to take on trades. It seems good to use.

    1 user thanked author for this post.
    avatar GraHal
    03/22/2020 at 11:50 AM #122894
    AutoStrategist

    Simple question: if we could know the spread amount in real time, wouldn’t you mind that this historical value is not available in backtesting?  

    Hello I’m back! If your wondering I stopped all trading for a while after I blew up my account 18 months ago with some careless manual trading.  Have been back running algos again since mid January and having been caught up in this spread fiasco and have some observations. To answer Nicholas I think the answer is a definite yes, if the spread could be checked before placing a trade that would be ideal.  Regarding the backtesting of spreads, I think the PRT interface as it stands could be improved, for futures trading across time ranges where (e.g. pre-market and market hours) I have to guesstimate some sort of average based on the number of trades in each time period.  Can the interface not be extended to allow different spreads (3 would be ideal) to be associated with different time ranges?

     

    03/22/2020 at 12:44 PM #122897
    GraHal

    Very practicable suggestion @AutoStrategist!

    Please propose it to PRT on the Form below … click the Suggestion Box in the left hand column.

    Use the email address that you are registered on the Account on which you access PRT and you might, just might get an acknowledgement of your suggestion?

    Sometimes I do get an acknowledgement, but mostly I don’t …  I’ve never figured out why I don’t??

    https://www.prorealtime.com/en/contact

     

    1 user thanked author for this post.
    avatar AutoStrategist
  • Author
    Posts
Viewing 15 posts - 31 through 45 (of 50 total)
1 2 3 4

Find exclusive trading pro-tools on ProRealCode Marketplace

Similar topics:
so no thrown out systems last night on fed’s decision – but no trades either

Create your free account now and post your request to benefit from the help of the community
Register or Login