TrashSim - OGame combat simulator

  • Hi there, thanks for your work...

    I found a small bug: when clicking on the second from right "+=-%" button, plunder is not adjusted.. I raided a bandit, plunder was 50%. Clicking the right button worked as expected (plunder is set to 100% automatically). I am using Firefox, if that should matter. You owe me 20m res :-)

  • Hi everyone,


    I've found some time again to invest into my OGame projects. I decided, apart from the MCO changes, to take another, deeper look into the simulation formula again.


    After a lot of testing and trying, I've found a fix and have written some new conclusions + started preparing an update. Which I'll hopefully release somewhere this month.


    To prepare for the update I also needed to update OGotcha to work with TrashSim in the background. In the future this wil enable me, and maybe you to, to verify in-game results with simulations, a bit more on that matter can be found in the OGotcha thread:

    OGotcha - Simple CR Converter for OGame v6+


    More info will come soon along with an update!


    Grtz,

    War

  • Dear Users,


    Version update 2.3.0.


    I've been wanting to update and fix the simulation formula for a long time now.

    But both my free time and my enthusiasm to update TrashSim were at a very low point.

    Support and responses from OGame haven't been the same as before the official release of v7, which makes me sad.

    I'm not going into this again, it makes me tired, for more info you can read this little thread: We're being left in the dark, a point of criticism, a complaint.


    Changelog:

    • Fixed: Overall simulation formula fixes. More about that below
    • Feature: A small testing back-office for me.


    So about the simulation formula fixes, before fixing the formula I first needed to run a lot of edge test cases to test a few hypotheses.

    Which took some time because I needed to work with new accounts every time because I couldn't downgrade my weapons technology.

    As some of you might know, the version 7 changes had some side effects on the simulation formula in-game (or there are just things we, or just I, didn't know before).

    Here are the conclusions I've come to when running all the tests, some of which are already know by some of you!


    1. First of all we have the "known by all" bounce effect

      We wanted to test if the bounce effect is also <= 1 on entities other than shields. After doing some tests on other types of defences and ships it seems to be the case. Everything below or equal to 1 bounces. But there's a different between the bounce % and the actual damage % on the shield. The bounce effect is calculated on attack / shield, whilst the damage % on the shield is floor( attack / shield ). -> check "percentage rounding" below for more info.

      ------------------------
    2. The damage effect

      Here we wanted to see if the damage percentage of the last ships spills to the hull points of the enemy entity. So if a ship has a damage percentage of 1.5%, the actual shield damage will be 1% per ship. So it takes 100 ships to kill the enemy entity. Now, will the 100th ship spill the 0.5% that's left on the hull? Well we haven't found a way to confirm this... sadly, because of the explosion flags :(

      -------------------------------------
    3. Damage percentage rounding

      In this test cases we're testing how the actual damage percentage is rounded. For instance if a ship has a damage percentage of 1.33333 or 1.66667 will it be ceiled to 2, floored to 1 or rounded -> above 1.5 to 2 and below to 1.

      In the tests we've noticed that everything below < 2 is rounded down to 1. So every damage percentage is floored to its nearest natural number. So 4.5654 to 4, 34.6363 to 34, ...

      Side note, something strange happens in case 24 & 27 (same set-up). The cruisers have the exact amount of power 50 * 600 = 30.000 which is the same as the shield shield being 30.000. Yet the LF in slot 2 are not able to kill the shield. So it appears the shield is still standing...

      This might not be the case for other defences, could only apply to shields. Case 29 is an example of that, there the PF its attack equals the HL its shield and the Recs in slot 2 do kill the HL. The recs's attack damage is below 1% just like the LF in the case 24 & 27.

      After some more testing on the it actually happens on every other enemy entity, so not only shields. Case 29 is an exeption as damage percentage is only calculated when the attack power is lower than the original shield value. If it's the same of more than it'll kill the whole shield. -> check "shield percentage fail" for more info on this.

      -----------------------------
    4. Shield percentage fail ?

      We've noticed something that we didn't know in forehand. If you check "percentage rounding" than you'll notice that if the ships (that don't bounce +1%) deal the exact amount of damage as the shields value it still stands. But this is only the case when the damage % is exactly the same as the actual shield damage %. So if the damage % is 2, than the shield damage % is also 2; if the damage % is 2.5, than the shield damage % is also 2. In the first case the shields still stands, so fodder (that would bounce) will fail and in the later the shield is destroyed, which means that fodder can kill the enemy. Which makes us believe that there is some sort of attack power spill, check "damage effect" for more info.

      In the Excel file these lines are marked in yellow.

      ----------------------
    5. Decimal rounding

      So here we wanted to test is the attacking power of, for instance, a small cargo is rounded. As they get an extra 0.5 for every weapons level. The test cases from 80 up until 99 confirm this, the attack power seems to be rounded down to it's nearest natural number. So 6.5 -> 6, 8.5 -> 8.


    You'll also find an Excel file (in a zip) containing all the tests I've done with the corresponding OGotcha report URL so you can get a more visual idea of the actual combat.

    I hope the column names speak for themselves, I you need more info on that matter, just let me know.

    The conclusions I've written above are also available in the Excel file.


    The small testing back-office:


    To easily confirm all my cases and test future cases I've also created a little back-office where I'm able to add OGotcha report ID's to my test cases.

    This enables me to run simulations of actual combats in OGame, including my 100 test cases :wink3:

    Because of this it'll be easier to test both changes to TrashSim or even OGame and compare them to see if everything is still correct!

    Here's a little example of how it looks and also to show you what the simulation formula fixes have done.


    First a screenshot of the old simulation formula:



    And this is what it now looks like :biggrin:

    So before 34 of the 100 test cases I've done fail and now they all succeed.

    At the bottom you'll see what it looks like when a test fails (in this case it's about a expedition battle which seem to be a bit weird in some cases)



    I hope this was informative enough, next I'll probably take a look at the different classes changes, wreckfield etc


    Grtz,

    Klaas

  • Hey guys.

    I think the simulator doesn't take into consideration 'General' class for calculating recyclers needed for debris after battle.

    Currently, the number of recyclers given to me by the simulator I divide it by 1.2 to get the real number.


    Just a thought! Hope I'm being usefull.

    Bye.

  • I'd like to contribute german translations. Thank you for updating to the latest patch, great work. How would you like me to do it?

  • Hi everyone,


    I'm looking for a full list of all the changes that happened and are currently not in TrashSim.

    Time is scarce currently and sadly I can't keep track of all the changes in the game as I'm also not playing the game at the moment.

    I am thinking about giving the simulator an update, so if anyone could provide me with the settings that are missing, then I'll look into it. :)


    Grtz,

    Klaas