Braindump of my proposed changes to GrandMasterPixel

Many of you have probably noticed (and quite a few players have commented) that the judging rate in GrandMasterPixel has become very slow in the last few months. This is frustrating since I know everyone likes to see how their pics do in the Arena quickly .. it’s pretty disheartening to submit you masterpiece, only to check two days later and find it hasn’t even been judged once !

As I’ve mentioned, this is to do with the having a larger ‘live’ pic population than when the game first started. My simple simulations suggest that with the current live pic population of about ~2000 pics, and about ~1000 battles daily (which was the case a few weeks ago), only about 15 % of pics should be missing out on getting at least one battle per day. The numbers will be a little different now, but you get the idea. Nonetheless … judging just seems too slow, so changes are afoot.

Proposal

For the latest GrandMasterPixel update, I increased the judging token recharge to one token every 15 mins (4X faster) .. but since many people will then also submit 4X faster, this won’t really solve the slow judging issue. The big change I have planned next is to increase the number of hitpoints for new pics (from 10 to 100) and the submission cost (from 10 to 40 submission credits per pic). I’ll probably also make it so pics lose 2 HP per lost battle, to prevent the pic population growing too much. I haven’t settled on these numbers 100% yet, but if my spreadsheet shenanigans are correct, the result of all this should be a faster judging rate, with a similar submission rate (about a maximum of two pics per day).

For the math/game design nerds: The other advantage will be that since all the numbers in the game are larger, it should be easier to tweak things more subtly in the future without moving away from whole numbers (eg, the smallest possible tweak to submission cost currently is 10 %, a change of +/- 1 to the total value of 10. A higher overall submission cost will make smaller percentage changes possible). Going to larger numbers kind of goes against my initial philosophy of keeping the numbers small to make the game more approachable to casual players … but now I’m thinking that this was a false premise, and it won’t be an issue for the majority of existing (or prospective) players.

Side effects

My biggest issue with making these changes is how to best deal with existing pics. Some people would very rightly complain if brand new pics suddenly got 10x more HP, allowing them to easily acquire more wins than an ‘old-world pic’ – so I will probably scale up the hitpoints for all existing pics by 10x (and current wins/loses by 5x), to match what would be equivalent if they had been submitted under the new system. This is probably not quite perfect, but it will have to do. At the same time, achievement thresholds related to numbers of wins will have to be multiplies by 5. In the process I just need to be very very careful to not completely mess up the database somehow :P

Bonus braindump !

While I’m at it – here’s another idea of something the might help speed up the perceived judging rate. I could prioritize judging for active players, so that if someone hasn’t judged for one week, their pics get flagged as “inactive” and don’t get included in battles as frequently as pics of active players. This would reward you guys that play every day, without requiring any major ‘economic’ changes immediately.

10 Responses to “Braindump of my proposed changes to GrandMasterPixel”

  1. DoomCabbage

    Sounds exciting! The thing with me is that I just love judging, but my creative juices just can’t flow fast enough to give me the tokens I need! Great work and can’t wait for more! A+!

    Reply
  2. Interesting! I’ve personally not been creating many more pictures but would love to judge some, yet I am greeted with the “you are a horrible filthy hoarder” message. I have quite a few pixel creations doing quite well in the arena at the moment and do not think I should have to add more to judge. Possibly there could be a system where as long as you have some living pixelart in the arena you can judge a certain amount each day even if your submission credits are full.

    Also, I miss many events that happen because of the limited space in the recent events scroll area.

    I would love love love it if I could see all my living pixelart on one page with hitpoints beneath them :) like a little art army doing battle.

    Anyway, still like the game, just not feeling inspired to create more art lately.
    Thanks!

    Reply
  3. 1. I feel the main problem is that people have to judge too few images, and that there is no growing price for submitting them. Level requirements, however, do increase.

    2. I have judged same images multiple times, so to me it feels like there aren’t many images at all!

    Reply
  4. I don’t know what the changes really did with the judging rate, but I noticed some things:

    1. Because the limit on tokens is half of that of credits, and tokens come only once every 15 minutes, I can only play two pictures a day if I keep a tight schedule; I feel this ruins the gameplay. Actually, if you were to remove the limit on tokens completely, or remove the tokens completely, while you are at it, things would become interesting again. There is, after all, already a limit on credits, so (a limit on) tokens seems redundant. Now, I only seem to be judging battles and waiting. At least, I would like to judge the amount I need at once, not in two runs.

    2. The number of wins and losses are all mixed up! Also, while some pics have battled five times, other pics have never battled. Maybe this is a little too random?

    3. If I am checking the battles of a specific pic, I can’t seem to go back the the previous page if that page is not the first page of the Pics pages. The Back button brings me back too far.

    Well, that’s it. I really hope that gameplay will continue to improve, and that the game will key on staying very cool!

    Reply
  5. Well, as disappointed as I seemed for my previous comment, I have learnt to live with the new rules. I feel I am in a position to contribute constructively, now. Okay, here goes:

    I really like the increase in judgments! I am actually surprised to see that the scoring has changed yet again, every time I check. So, this could just be as much judging as is needed for the battling to be fun. Good job, there!

    I can imagine that the game economy is very interesting, but I feel the interface is still ‘lagging’. I mean, of course, points mentioned in comments on previous posts, especially on the usability of web-based screens and, sorry to say so, the main screen; I feel that the addition of functions is not reflected in anything else than the notifications, and those are not the primary ‘selling point’. Since this is beta, I am attracted to seeing the game grow. Not just by playing, but by ‘checking it out’.

    Final note: I still feel that the restriction tokens is a bit harsh; that is, the last two pics I submitted were the second I drew; the first ones never made submission because I got a ‘better’ idea. Being able to save pics (and reuse after editing!) would be ingenious feature (and a restriction on submitting identical or very similar pics, like Link).

    Reply
  6. Cheater spotted, you might check security! SMS! by NF45 can’t be a submitted pic, although it displays the only annoying bug in GMP.

    Greetz!

    Reply
  7. It seems there is no cheater after all (spotted another one of the same weird thing), but the ‘ability’ of the game to submit whatever is on the screen.

    Reply
  8. Thanks for your feedback RMC. Lots of things to address here, but the short answer is – I basically agree with everything you are saying :) Most of the features are things I’d like to add but haven’t yet due to time constraints. The last few months I focussed on dealing with the judging rate issue in my spare time, which resulted in the scoring changes recently implemented. Not everyone has been happy with that, but I’m hoping on balance it makes submissions more satisfying … the number one complaint from players was “My pics aren’t getting judged”.

    Regarding:

    3.1 – the pricing for submission vs. amount of judging required is geared such that the game shouldn’t run out of pics. Any pricing change needs to ensure that every player will effectively replace the pics that they kill through judging. This is the basis for many of the annoy aspects in the scoring.

    3.2 – I’ve noticed the same pic coming up many times also (excluding lame dupes like “Link”). Others, like yourself, have also noted it. I dug into the code a while back to check that there was no significant bias in the pseudo-random numbers used to generate battles – as far as I can tell they are truly random, it’s just a psychological effect of recognizing familiar pics if they come up again by chance. I’ve been thinking of implementing a check to ensure that the same judge doesn’t seem the same pic too frequently .. it’s on the TODO list somewhere (along with prioritizing battles for pics of users who have recently played).

    4.1 – The rate limit on tokens (and hence effective limitation on submission credits earned per day) is so that people take care to submit good pics, rather than just submit spammy pics in order to level up. That said, for those wanting to submit a steady stream of high quality pics like yourself, I can see how this rate limiting is annoying. The original design was geared so that players could spend 5-10 mins per day judging, in two sessions (eg morning and afternoon) and get enough credits to submit two pics (again, one in the morning, one in the afternoon). It is a deliberate decision to prevent players from being able to play too much in one day … basically since I don’t have time to play much more than that, and I needed to level the playing field against kids with too much time on their hands :). The new scoring scheme breaks this a little – I guess it would take at least 12 – 15 min of judging, across two sessions, in order to submit only one pic per day. On re-examination, this is a little onerous and unfun. Maybe raising the judging token cap to 40 (or 60 ?) would help somewhat ? This way judging to earn enough credits for one submission could be done in a single session. One way to speed this up would also be to make the judging inteface (“Arena”) faster .. more about that later.

    4.2 – It would be good if battles were more evenly distributed between pics .. but how exactly to do this, and do it right, isn’t as easy as it first seems. If I were to give battle priority to pics with fewer battles to try and even it up, new pics would tend to get lots of action, neglecting older successful pics. It’s always interesting how ‘completely fair’ distribution of battles ends up looking unfair though – maybe if I did prioritize battles for pics with zero wins+loses, things would *appear* more balanced (even though they wouldn’t be). An alternative way I think is to simply give priority to battles for pics of users who have recently played – this won’t even up the battles, but it will mean active players get their pics judged more quickly, which should be more satisfying.

    4.3 – There is a “previous page” and “next page” link on the fight history page that work, but the Android “Back” button will back out of the inbuilt browser entirely. I know this breaks Android conventions, and I should fix it … the idea was to make the inbuilt browser not so browser-ish so that people didn’t expect it to behave that way … but obviously that plan hasn’t worked out.

    And that brings me to the interface.

    I agree that the interface could do with a lot of work. As long as there isn’t need to dramatically overhaul the scoring system again (given the few extra tweaks discussed above), I can focus on UI improvements and new features again (eg, the often requested local save/load feature). The original idea of using an embedded web interface was that I could easily port GMP to multiple platforms .. namely Facebook and iPhone. The amount of time it’s taken getting the game to where it is, I think those ports are a long way off. Also, while the iPhone App Store has become more transparent in recent months, I’m not a huge fan of the way Apple manages their platform, and I’m not sure if I want to risk wasting time on an iPhone version only to have it rejected because some Apple drone spotted some ‘offensive’ user generated content. I’d rather let the players decide what they like, not Apple. So, the upshot of this is that either the web interface needs to be dramatically improved (tough, but possible), or I start building a native Android interface that has the same sort of functionality (my preferred option). A native interface should be really snappy particularly if lots of caching of pics is used. It will be sad to throw away all the work I’ve done on the web interface, but the more I think about it, the more I’m convinced it’s the best way forward.

    The other idea I’ve been toying with, but have avoided mentioning lest people start demanding it faster than I can do it, is to open source the whole kit-n-kaboole. This would allow others to help fix bugs, add features and even start their own servers with different rules etc (I’d add a global server list to the client). In reality, for project like this, I doubt many developers would contribute much, and there aren’t really enough players to warrant multiple servers at this time. The main thing holding me back is that the code really needs a cleanup to remove embarrassing bits before it is seen by anyone with a skerrick of programming experience.

    Reply
  9. Regarding the ‘cheaters’ (pics showing the palette, for instance) .. I’m embarrassed to say this is a known bug/exploit, that I was hoping no one would find (wishful thinking, I know :P). I can be fixed, and even those pics that show the bug can be reverted on the server to whatever the player *actually* drew, but I haven’t got around to fixing it yet since it’s not a trivial one line bugfix, but will take a small pile of extra code. As far as I can tell ‘NextTarget’ was the first to submit a pic demonstrating the bug … but until now no one has actually told me about it.

    Thanks :)

    Reply

Leave a Reply

XHTML: You can use these tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>