Help me collect statistics.

Discussion in 'Hide and Seek' started by dalahast, Aug 7, 2013.

  1. uH8R

    uH8R The Hive's Batman

    Joined:
    Aug 1, 2013
    Messages:
    140
    Ratings:
    +135
    Happy to help! 15 games, I know you wanted 20 but I was tired :(

    Code:
    villa 8
    space 3
    townsquare 10
    animalvillage 19
    space 11
    animalvillage 13
    villa 5
    animalvillage 13
    hotel 3
    space 4
    townsquare 3
    hotel 3
    townsquare 7
    townsquare 14
    villa 3 
    #21
  2. dalahast

    dalahast Well Known Bee

    Joined:
    Jun 29, 2013
    Messages:
    221
    Ratings:
    +130
    We're up to 161 now. 39 (or more) left before I analyze it.
    #22
  3. Blazing_Leaf

    Blazing_Leaf V.I.BEE

    Joined:
    Mar 3, 2013
    Messages:
    1,149
    Ratings:
    +668
    I'm playing some Atm... Have 4, On my Fifth...

    EDIT:
    Code:
    space: 17
    townsquare: 6
    space: 7
    villa: 21
    space: 2
    hotel: 6
    villa: 5
    townsquare: 9
    villa: 5
    space: 3
    villa: 17
    hotel: 1
    hotel: 4
    townsquare: 5
    space: 5
    hotel: 0
    hotel: 1
    townsquare: 6
    space: 3
    villa: 12
    animalvillage: 13
    animalvillage: 14
    #23
  4. ZipZero

    ZipZero Regular Bee

    Joined:
    Aug 8, 2013
    Messages:
    13
    Ratings:
    +1
    Here's 39:
    Code:
    villa 8
    villa 12
    animalvillage 6
    townsquare 16
    hotel 0
    townsquare 6
    animalvillage 11
    animalvillage 11
    animalvillage 5
    space 6
    animalvillage 12
    townsquare 2
    villa 10
    animalvillage 11
    townsquare 12
    villa 14
    villa 6
    villa 8
    townsquare 13
    hotel 4
    hotel 7
    animalvillage 21
    animalvillage 12
    space 9
    hotel 1
    townsquare 7
    villa 13
    space 10
    villa 8
    villa 7
    hotel 1
    hotel 5
    townsquare 5
    townsquare 12
    townsquare 19
    animalvillage 12
    hotel 1
    animalvillage 16
    animalvillage 15
    #24
  5. dalahast

    dalahast Well Known Bee

    Joined:
    Jun 29, 2013
    Messages:
    221
    Ratings:
    +130
    That's 222 games. I'll be editing this post with results in a moment, stay tuned.

    Edit: Results are in. Lots of interesting stuff this time around. We've got new maps to examine this time, and more detailed data. Thanks to all the players who recorded their game results for me, I had 222 games to analyze this time, more than any of my previous experiments.
    • I once again checked to make sure that the maps were showing up randomly, as they're supposed to; lots of people were keeping track of the number of times they got the same map in a row in some of the lobbies. The maps still appear to be random, each showing up about 20% of the time (χ² = 2.41, p = 0.66).
    • The average number of hiders remaining on each map is as follows: 12.57 on Animal Village, 2.79 on Hotel, 6.90 on Space, 8.75 on Town Square, and 6.50 on Villa. The average for all maps combined is 7.68 (σ = 5.23).
    • Not surprisingly, the average number of hiders varies widely by map (a one-way analysis of variance was significant: F(4,217) = 30.07, p < 0.001). In particular, Animal Village tends to have many more hiders left than the other maps.
    • The seekers are consistently performing worse and worse. This time around, seekers won only 3.6% of all games (that's 8 out of 222). Of those 8 games, 6 were on Hotel, while 2 were on Villa. No game on Animal Village, Space, or Town Square was won by seekers. Notably, no game on Animal Village finished with less than 5 hiders.
    • The largest number of hiders remaining at the end of any game was 21. Curiously, the most common number of hiders remaining at the end of a game is 6 (12% of all games ended with 6 hiders). On the other hand, roughly half of all games finish with 7 or more hiders.
    The extreme unfairness of Animal Village, Space, and Town Square has been confirmed. Animal Village, in particular, was found to have significantly more hiders remaining on average when compared to other maps. Hotel is no longer alone when it comes to maps that the seekers can actually win on, having now been joined by Villa. Still, the overall win rate of seekers is lower than it has ever been estimated to be in the past, with less than 4 games out of 100 being won by seekers. Approximately half of all games played finish with at least 7 hiders remaining, and sometimes as many as 21.

    I'll be drawing up plans for the next experiment sometime soon, so be on the lookout for that. It will be posted in this thread, and I'll update the OP accordingly.
    #25
  6. Xtrim2008

    Xtrim2008 V.I.BEE Emerald Member

    Joined:
    Dec 30, 2012
    Messages:
    2,266
    Ratings:
    +1,555
    Should I post my results anyway?
    #26
  7. ZipZero

    ZipZero Regular Bee

    Joined:
    Aug 8, 2013
    Messages:
    13
    Ratings:
    +1
    The results are quite interesting, I never really noticed how few times the seekers win before recording the data. I would like to say, though, that in many of the games where only a few of the hiders were found, it was because the seekers were just terrible at seeking. For instance, in one game on Villa, the seeker hadn't found anyone by 4 minutes into the game, and quit in the last remaining minute. I was randomly assigned to be the new seeker, and found several hiders in just that short time period. I was also spawnkilled by a swarm of angry blocks, but that's a separate issue.
    Another note about the apparent lack of skill within the seeker ranks is that whenever I hide inside a room, particularly in Hotel and TownSquare, most of the seekers will just walk into the room and then walk out without checking any of the blocks, or checking random blocks in the wall which couldn't be a hider. Many even don't go into any rooms at all. I've played several games in TownSquare without seeing a single seeker, despite being in a fairly conspicuous place in the middle of a room. I'm not saying the lack of seeker competence is the main reason for hiders winning nearly every game, but it certainly plays a part in it.

    P.S. To get accurate results, I had to play every game I went into, even if I hated the map. Including Space. I hate Space.
    #27
  8. pompeyftw

    pompeyftw V.I.BEE

    Joined:
    May 16, 2013
    Messages:
    1,577
    Ratings:
    +1,201
    I would have posted more but my mum was all like WHY DA HELL YOU STILL ON DA COMPUTER GET TO BED NOAW so yeah. I shall collect more results today.
    #28
  9. dalahast

    dalahast Well Known Bee

    Joined:
    Jun 29, 2013
    Messages:
    221
    Ratings:
    +130
    Indeed, I've played a good number of games with incompetent seekers, too. Nonetheless, it's important to include games with them in these experiments, in order to accurately reflect the conditions on the server. With the seeker win rate at a dismally-low 3.6%, I'm sure the maps could be made a little bit easier so that the incompetent seekers could actually get something done, while still providing a challenge for the more experienced players.
    #29
  10. Blazing_Leaf

    Blazing_Leaf V.I.BEE

    Joined:
    Mar 3, 2013
    Messages:
    1,149
    Ratings:
    +668
    May I say... the game with 21 hiders left on villa had a seeker who didn't know how to sprint...
    #30
  11. dalahast

    dalahast Well Known Bee

    Joined:
    Jun 29, 2013
    Messages:
    221
    Ratings:
    +130
    Mhm. As I mentioned above, it is important to record that kind of thing nonetheless, since it happens often enough to affect the game significantly.
    #31
  12. ZipZero

    ZipZero Regular Bee

    Joined:
    Aug 8, 2013
    Messages:
    13
    Ratings:
    +1
    It looks like the recent removal of the rocket ladder from Space was a good thing, I just played a game there where there was only one hider left.
    #32
  13. dalahast

    dalahast Well Known Bee

    Joined:
    Jun 29, 2013
    Messages:
    221
    Ratings:
    +130
    Once the next map(s) are released, I plan to repeat this last experiment and compare the results of it to the results obtained before. This way, not only can we tell whether or not the new maps are better than the existing ones, but we can also determine whether any changes made to the existing maps were effective in giving the seekers a chance at winning.
    #33
  14. ejegun

    ejegun Regular Bee

    Joined:
    Mar 23, 2013
    Messages:
    34
    Ratings:
    +20
    Wow this is really interesting. Thank you :D
    #34
  15. dalahast

    dalahast Well Known Bee

    Joined:
    Jun 29, 2013
    Messages:
    221
    Ratings:
    +130
    Heh, I'm glad you found it interesting. I definitely find it interesting too, which is why I started doing it in the first place. It's uncovered a lot of things that aren't obvious from just playing the game, not the least of which is the apparent impossibility of most maps for the seekers. If you compare this to the surveys some other users did on here a while back, you'll also see that whether or not people like a map may have almost nothing to do with how fair it is: Town Square beat Hotel in the approval-rating polls another user did, despite Town Square having virtually no seeker wins, ever.
    #35
  16. Queekusme

    Queekusme Regular Bee

    Joined:
    Jul 19, 2013
    Messages:
    34
    Ratings:
    +1
    There's a reason for Animal Village and that's because blocks can get out of the map... Once they fix this exploit I'm sure hider survival rates will lower
    #36
  17. ejegun

    ejegun Regular Bee

    Joined:
    Mar 23, 2013
    Messages:
    34
    Ratings:
    +20
    There's this nice button called "reply" it does what I've done ^here.
    I think the amount of times each map gets played is pretty random. But technically, computers cant do random things.
    #37
  18. dalahast

    dalahast Well Known Bee

    Joined:
    Jun 29, 2013
    Messages:
    221
    Ratings:
    +130
    There is indeed a "reply" button. As a rule, I do not use it if I'm replying to the person immediately above me, as it's a waste of space to reproduce their entire post right below the original, IMO. If I'm not replying to the most recent post in the thread, then I'll hit it.

    That's true; technically computers can only be pseudorandom—but if pseudorandom is random enough that it doesn't produce a statistically-significant difference between what is expected (every map showing up 20% of the time) and what actually happens (by random chance, it is entirely possible and reasonable for one map to maybe show up 21.9% of the time and another only 18.1% of the time in experimental data), then I cannot conclude that the maps aren't chosen randomly.
    #38
  19. ejegun

    ejegun Regular Bee

    Joined:
    Mar 23, 2013
    Messages:
    34
    Ratings:
    +20
    Yup. <--Short reply because I'm tired and will go to sleep in a sec.
    #39
  20. Queekusme

    Queekusme Regular Bee

    Joined:
    Jul 19, 2013
    Messages:
    34
    Ratings:
    +1
    What might be needed is something like a modified random...

    Theory:

    Computer picks a number... that number corresponds to a map.
    If the number hasn't been picked then it flags it as picked and it then forwards that out so it can be used...
    Else it picks again (and again until an un-chosen number is found)
    Repeat for each map until all are picked...
    Reset counters.

    That way, all maps get an equal play amount, but it's still random in which one appears in a given repetition.
    Also doing this, you can re-seed the random each time meaning that it's even more randomly chosen...
    #40

Share This Page