The new forums will be named Coin Return (based on the most recent vote)! You can check on the status and timeline of the transition to the new forums here.
The Guiding Principles and New Rules document is now in effect.

[TRENCHES] Thursday, August 7, 2014 - Sherblocked

GethGeth LegionPerseus VeilRegistered User, Moderator, Penny Arcade Staff, Vanilla Staff vanilla
edited August 2014 in The Penny Arcade Hub
Sherblocked


Sherblocked
http://trenchescomic.com/comic/post/sherblocked

One Item rules it all

Anonymous

Recently, Its been my task to acquire all available items in the game.

At first, it was not all bad… Until they released an update.

This particular update was not really pleasant because there was one item that can only be obtained by chance in a lottery game.

Easy, right? Nah.

With my sanity still intact, I asked the dev what are the chances of getting it and they said “oh, about 0.18%”.

With a bad poker face I said “Okay.”

Throughout the rest of the shift like a maniac, going back and forth through the lottery tapping away with all the energy I can muster up just to obtain one item and still no luck.

Finally my boss, who’d been watching my ordeal, understood my position and told me that its alright if I can’t get it since we were on a one day build.

I consider myself lucky that I wasn’t stuck with the other team who had to get all 1000+ items which are all randomly generated all in one day.


Geth on

Posts

  • killer.zombierobotkiller.zombierobot Registered User new member
    No. I refuse to believe this Tales, and if it is indeed true, all nonsense you had to deal with is your own fault. There's a concept called a unit test. You might have an automated test that determines, through repeated random execution, the probability programmed in are the actual drop chances from that particular algorithm. If your entire testing concept is so incredibly flawed that you either a) don't trust your RNG or b) don't trust the interface to your RNG system and c) can't move past that to just assign the random drop to the test character, your entire team is screwed and deserves to burn in the seventh pit of hell.

  • RMS OceanicRMS Oceanic Registered User regular
    Unit Tests are great! Thing is it takes time to develop them, and maybe these guys didn't have time budgeted for that.

  • HollywoodRPGHollywoodRPG Registered User new member
    The tales are still somewhat interesting, but I have lost all interest in the comic.

  • plki76plki76 Redmond, WARegistered User regular
    No. I refuse to believe this Tales, and if it is indeed true, all nonsense you had to deal with is your own fault. There's a concept called a unit test. You might have an automated test that determines, through repeated random execution, the probability programmed in are the actual drop chances from that particular algorithm. If your entire testing concept is so incredibly flawed that you either a) don't trust your RNG or b) don't trust the interface to your RNG system and c) can't move past that to just assign the random drop to the test character, your entire team is screwed and deserves to burn in the seventh pit of hell.

    Exactly. Why would someone manually try to reproduce something that has less than a 1% chance to occur? That's absurd.

    To those who said there may not have been time to automate this, it shouldn't take more than half a day to write something like:

    int attempts=0
    Item item;

    while(attempts<MAXATTEMPTS)
    {
    attempts++;
    item=TryLottery();
    if (item == itemwearelookingfor) break;
    log(item); //Can use this data to verify the percentages of items and possibly save another tester some time while you're at it
    }

    if (attempts == MAXATTEMPTS) printf("Failed to get item.");
    else printf("Got item in %n attempts", &attempts);

  • marsiliesmarsilies Registered User regular
    No. I refuse to believe this Tales, and if it is indeed true, all nonsense you had to deal with is your own fault. There's a concept called a unit test. You might have an automated test that determines, through repeated random execution, the probability programmed in are the actual drop chances from that particular algorithm. If your entire testing concept is so incredibly flawed that you either a) don't trust your RNG or b) don't trust the interface to your RNG system and c) can't move past that to just assign the random drop to the test character, your entire team is screwed and deserves to burn in the seventh pit of hell.
    I don't think the testing was to test the probability, but to test if the game broke when acquiring an item, or when acquiring all items, through normal gameplay.

    But yes, automating the process some way would've probably have been better.
    http://trenchescomic.com/comic/post/neologism

  • tardigradetardigrade Registered User new member
    I really don't understand how Cora, who is supposed to have a managerial role, is somehow consoling an employee who has just admitted to creating a clone to compete with her game, based on her own father's intellectual property. Are the writer(s) just not keeping track?

Sign In or Register to comment.