Radio Drop Rates and what do we take away? (NEW: 15x double hero-token call)

2

Comments

  • jimmorrison369jimmorrison369 Member Posts: 2,065
    Putchuco said:


    I must admit I thought the 15x standard call would've been even better than it is, but it is quite close to the standard 5x call. Surprisingly.

    Well, maybe it is only .35 tokens per radio less.
    But you also can choose more specific survivors. So all in all, it is significantly better to do the 15 call, they are by far comparable.

  • enigma12enigma12 Member Posts: 248
    @Putchuco Thanks for the great work you put in! It’s greatly appreciated. There are a couple of us in our guild who seem to be the odd ones out and never have any luck with the game. But others seem to be extremely lucky, even though I was blatantly told ‘luck’ doesn’t exists. Lol Whatever you call it, some of us are not able to relate to the stats but there is always hope. Thanks again for everything!
    Putchuco
  • PutchucoPutchuco Member Posts: 1,289
    @enigma12 thx much for the kind words!
    I will share the appreciation with @jimmorrison369 who is working with me on these calculations.

    In terms of luck or lack thereof: yes, there is a large standard deviation in these calls, ie the results vary wildly around the expected average results.
    I really want to calculate these standard deviations to give us a better feeling of what ‘range’ of results we are to expect. I have a feeling the deviation is not the same on all the calls, some are closer to the average than others.
    But first, I need to brush up on my math behind std deviations... lol

    Pls let us know if you can work with the provided tables, if you need it represented differently, or if you are missing numbers, something isn’t quite clear, anything.
    We want to present this as ‘easy to the eye’ as possible, no math degree should be required to read them.
    If you would like to join a drama-free and fun group of enthusiasts,
    contact us via email at [email protected]
  • JackBauerJackBauer Member Posts: 1,550
    I am feeling lost. So what do we take away from it ?
    -----------------
    Remember to use your charge abilities
  • PutchucoPutchuco Member Posts: 1,289
    edited February 2018
    @JackBauer I believe the most important thing to take away here is this overview:



    It shows us the summary of all the calls we have looked into so far (more to come),
    and it can be used to
    a) compare the various radio calls to each other (‘which is better, 15x or 5x?’)
    b) to get a feel what is the expected average outcome of spending say 1000 radios on a call. For example: I may need 750 token for a hunter upgrade, so i know where to look to get them.

    The new 20x Special Hero call this week (for Carol) gives us way more Carol token than the ‘normal’ 10x special hero call, in average.

    Careful: the numbers are the average numbers, actual radio calls vary wildly around these values, ie you wont hit these numbers reliably when spending only a few radios.

    Does this make sense?
    If you would like to join a drama-free and fun group of enthusiasts,
    contact us via email at [email protected]
  • MadPuppyMadPuppy Member Posts: 2,831
    Consice version: the odds are excellent but there's still a big chance of getting low results. Keep your hopes realistic to avoid unneccessary disappointment.
    TJS
  • dylanarenadylanarena Member Posts: 78
    edited February 2018
    @Putchuco and @jimmorrison369, I wrote a Python script on a flight the other day to simulate the various flavors of radio calls. (That way I could be sure I wasn't fumbling the probability calculations.)

    If either of you is familiar with Python, I'd be happy to send you the file to supplement the work you're doing with Excel. Here's a sample of the output when you simulate spending 9M phones (I choose multiples of 900 so I can spend all phones on 1, 5, 15, or 25 calls):

    compare_expected_returns(9000000)
    1-phone calls:
    0.40 hero tokens and
    3.41 survivor tokens per phone
    ---------------------------------------
    5-phone calls:
    0.57 hero tokens and
    4.44 survivor tokens per phone
    ---------------------------------------
    15-phone calls:
    0.76 hero tokens and
    4.59 survivor tokens per phone
    ---------------------------------------
    25-phone calls:
    1.75 hero tokens and
    1.36 survivor tokens per phone
    ---------------------------------------

    These results consider all Heroes to be exchangeable and all Survivors to be exchangeable, and the way I handled the "priority" issue (i.e., how to choose among three cards when some are Heroes and others are Survivors) is to declare a constant multiple for "Hero premium". In this run, I've used 3, which means I'd choose a 32 Survivor over a 10 Hero and a 128 Survivor over a 32 Hero.

    The tl;dr for me is that the 15 calls offer the best yield for Survivor tokens and the 25's offer the best yield for Hero tokens. I haven't added in the 40 calls yet, but the way I've written the functions it'll be straightforward to add that logic.
    Putchucojimmorrison369JackBauer
  • PutchucoPutchuco Member Posts: 1,289
    @dylanarena that sounds awesome!
    Yes, pls send them, we will certainly look at them as one of my guildmates just started doing something very similar. This might help to jumpstart these efforts.
    If you would like to join a drama-free and fun group of enthusiasts,
    contact us via email at [email protected]
  • jimmorrison369jimmorrison369 Member Posts: 2,065
    edited February 2018


    compare_expected_returns(9000000)
    1-phone calls:
    0.40 hero tokens and
    3.41 survivor tokens per phone
    ---------------------------------------
    5-phone calls:
    0.57 hero tokens and
    4.44 survivor tokens per phone
    ---------------------------------------
    15-phone calls:
    0.76 hero tokens and
    4.59 survivor tokens per phone
    ---------------------------------------
    25-phone calls:
    1.75 hero tokens and
    1.36 survivor tokens per phone
    ---------------------------------------

    Wow, another confirmation we are using the right formula.

    Awesome @dylanarena
    Above is exactly what I've found using formulas

    Except the 25 call, but that is a little different route to take. Also one can question if you roll a guaranteed hero card, or dismiss any outcome that has no hero in there. Not sure how to approach it.

    I've done a little sheet to calculate tokens per radio.
    This is what I came up with

    1 radio call = 4.00 3.8 tokens per radio
    5 radio call = 5.26 5.07 tokens per radio
    15 radio call = 5.44 tokens per radio
    25 radio call = 3.27 tokens per radio
    10 radio call = 2.15 2.07 tokens per radio

    Edit, thought uncommon survivor was 10 tokens. lol


    Second Third edit.

    1 radio call = 3.81 tokens per radio
    5 radio call = 5.01 tokens per radio
    15 radio call = 5.35 tokens per radio
    25 radio call = 3.21 tokens per radio
    10 radio call = 2.58 tokens per radio.

  • dylanarenadylanarena Member Posts: 78
    @jimmorrison369, for the 25 call I passed in probabilities that guaranteed a hero for the first card daw (with the rarities listed in the info sheet—I collapse specific Heroes and Survivor classes for the summaries, but my simulations do draw, e.g., 10 Abe or 32 Scout or whatever) and then drew normally for the remaining two draws.

    @Putchuco, I will PM you the Python script (perhaps right after I finish implementing the 40-call logic!).
  • enigma12enigma12 Member Posts: 248
    @Putchuco You really are a gem and such a nice dude! I will look through your work and consult with my guild and if there are any ideas that may help, I will let you know. We all appreciate the work you put in!
  • PutchucoPutchuco Member Posts: 1,289
    My pleasure @enigma12 :)
    If you would like to join a drama-free and fun group of enthusiasts,
    contact us via email at [email protected]
  • PutchucoPutchuco Member Posts: 1,289
    Here's a little experiment:
    I combined Carol's special hero chances with her base chances being a Rare Hero.

    You will see that the difference really is minimal:
    - combined chances (as shown below): 1138 Carol, 17 per Hero, 405 per survivor class
    - not combined (as shown in first post): 1128 Carol, 16 per Hero, 408 per survivor class




    I will therefore not go through the trouble of combining those.

    The same is true to a certain degree for very desirable heroes like Abe and Sasha.
    Many players will prefer a 128 Abraham over a 256 Jerry as an example,
    but breaking this down for each and every case would not add much value to the topic.
    If you would like to join a drama-free and fun group of enthusiasts,
    contact us via email at [email protected]
  • ShadowaceAzShadowaceAz Member Posts: 3,402
    Putchuco said:


    The same is true to a certain degree for very desirable heroes like Abe and Sasha.
    Many players will prefer a 128 Abraham over a 256 Jerry as an example,
    but breaking this down for each and every case would not add much value to the topic.

    Really needs to be a Sad added to the like agree disagree lol selection..
    because that just makes me want to


    Descensus in cuniculi cavum
    Vir prudens non contra ventum mingit
    Concordia

  • PutchucoPutchuco Member Posts: 1,289
    Don‘t cry, there, there @ShadowaceAz
    Now tell us what made you so upset
    If you would like to join a drama-free and fun group of enthusiasts,
    contact us via email at [email protected]
    TJSDodkongbigbeano
  • ShadowaceAzShadowaceAz Member Posts: 3,402
    Me be needing well over 200 Carol tokens to get the ‘Spear of Destiny’ and your latest informs me my 100 pocket radios will not be strong enough to get a wide band signal out. I may as well be using cans and strings...
    Descensus in cuniculi cavum
    Vir prudens non contra ventum mingit
    Concordia

  • PutchucoPutchuco Member Posts: 1,289
    Well, you could get lucky, a 128 Carol is almost at 10% odds
    If you would like to join a drama-free and fun group of enthusiasts,
    contact us via email at [email protected]
  • Amcwdfan1Amcwdfan1 Member Posts: 1
    Does anyone know how I find out how many phones I have collected?
  • ShadowaceAzShadowaceAz Member Posts: 3,402
    Top left of you base page shows the number of phones. If it’s still saying Free, you make that free call and the number reappears.
    Or open the radio tent like you’re going to call, it shows how many you have available.

    That help?
    Descensus in cuniculi cavum
    Vir prudens non contra ventum mingit
    Concordia

    David_H79
  • ShadowaceAzShadowaceAz Member Posts: 3,402



    Phone count displayed.
    Descensus in cuniculi cavum
    Vir prudens non contra ventum mingit
    Concordia

    David_H79
  • dylanarenadylanarena Member Posts: 78
    Update: I've added the 40-phone logic to my simulation script (two versions: reject 32-token survivors and reject 32-token survivors *and* heroes) and augmented the output to show a bit more. My takeaways are that (a) Hero Fest gives the most bang for the buck (albeit with way more hero tokens and slightly fewer survivor tokens than the 15 call) and (b) in the long run, it pays to keep your Hero Fest bar at 128 tokens of any type rather than to accept even 32-token heroes.


    Here's a run with 6M phones (because 600 divides cleanly by 1, 5, 15, 25, and 40):


    compare_expected_returns(6000000)
    ---------------------------------------
    1-PHONE CALLS
    ---------------------------------------
    0.4 hero tokens and
    3.41 survivor tokens per phone
    -----
    3.81 total tokens per phone
    ---------------------------------------
    5-PHONE CALLS
    ---------------------------------------
    0.567 hero tokens and
    4.439 survivor tokens per phone
    -----
    5.006 total tokens per phone
    ---------------------------------------
    15-PHONE CALLS
    ---------------------------------------
    0.754 hero tokens and
    4.59 survivor tokens per phone
    -----
    5.343 total tokens per phone
    ---------------------------------------
    25-PHONE CALLS
    ---------------------------------------
    1.749 hero tokens and
    1.357 survivor tokens per phone
    -----
    3.106 total tokens per phone
    ---------------------------------------
    40-PHONE CALLS (REJECTING 32-TOKEN SURVIVORS)
    ---------------------------------------
    1.497 hero tokens and
    4.196 survivor tokens per phone
    -----
    5.693 total tokens per phone
    ---------------------------------------
    40-PHONE CALLS (REJECTING 32-TOKEN ANYTHING)
    ---------------------------------------
    1.307 hero tokens and
    4.651 survivor tokens per phone
    -----
    5.957 total tokens per phone
    ---------------------------------------


    jimmorrison369MadPuppymsapaevShadowaceAz
  • PutchucoPutchuco Member Posts: 1,289
    edited March 2018
    @dylanarena awesome run there, this certainly helps verifying the calculations!

    I took a stab at the 25x guaranteed hero call:



    Pretty close to your numbers.

    EDIT: Fixed error
    If you would like to join a drama-free and fun group of enthusiasts,
    contact us via email at [email protected]
  • PutchucoPutchuco Member Posts: 1,289
    edited March 2018
    @dylanarena here's a quick comparison of your scripted results vs the formula-based calculations:





    If you would like to join a drama-free and fun group of enthusiasts,
    contact us via email at [email protected]
    msapaev
  • dylanarenadylanarena Member Posts: 78
    Thanks, @Putchuco! That seems like a good confirmation that we're doing our respective analyses more or less correctly. :)
    jimmorrison369Putchuco
  • PutchucoPutchuco Member Posts: 1,289
    Absolutely @dylanarena , this is great!
    It’s so easy to get lost in excel formulas, gotta double check them ;)
    If you would like to join a drama-free and fun group of enthusiasts,
    contact us via email at [email protected]
  • msapaevmsapaev Member Posts: 480
    edited March 2018
    @Putchuco, in the "hero token per hero per 1000 radio" column of your average turn-outs table, the numbers reflect tokens of an "average hero", don't they? How do these numbers fit heroes' rarity, for example, Carl vs Jesus?

    For example, for 15x call, it's 34.5. Should I adjust this number according to the hero's rarity — decrease for "epic" Carl and increase for "common" Jesus?
  • jimmorrison369jimmorrison369 Member Posts: 2,065
    msapaev said:

    ...the numbers reflect tokens of an "average hero", don't they?— decrease for "epic" Carl and increase for "common" Jesus?

    Basically yes, it does depend somewhat if you pick carl tokens over jezus tokens.
    But for all intents and purposes, yes.
    msapaev
  • PutchucoPutchuco Member Posts: 1,289
    @msapaev what @jimmorrison369 said. We wanted to keeep the formulas at a reasonable length. We could split it for each hero or hero rarity but would be truly mindblowing.

    If there is a good reason to do so i’d be happy to.
    If you would like to join a drama-free and fun group of enthusiasts,
    contact us via email at [email protected]
    msapaev
  • msapaevmsapaev Member Posts: 480
    edited March 2018
    @Putchuco, I don't think it's necessary. I think the understanding that you have more chances to get 34.5 Jesus's tokens than 34.5 Carl's tokens is enough. Thanks a lot for clarifying and for your (and @jimmorrison369 and @dylanarena ) awesome and useful analysis!
  • DrunkenDrunken Member Posts: 1,726
    @Putchuco

    I want it split all the way down. What are my odds of getting Michonne tokens at 512 on a non double event?
    Putchuco
Sign In or Register to comment.