this is a test for the ##vote: nisani teamliquid parser
+ Show Spoiler [test] +
##vote: quatol
##unvote
##vote: gonzaw
Forum Index > TL Mafia |
Nisani201
United States1400 Posts
this is a test for the ##vote: nisani teamliquid parser + Show Spoiler [test] + ##vote: quatol ##unvote ##vote: gonzaw | ||
gonzaw
Uruguay4911 Posts
If you just read the final posts and go backwards you'd probably get the most info and you could ignore the initial posts (except maybe the op). | ||
GreYMisT
United States6736 Posts
| ||
gonzaw
Uruguay4911 Posts
On November 28 2012 14:14 GreYMisT wrote: How would you handle replacements? In the system a player is just a username. I just replace one of them with another one Or rather, set the previous one as "replaced" and create a new one and give him the other guy's role/alignment/etc. The host would just PM the bot "X is a replacement" and people that don't vote are automatically replaced when they fail to vote (if multiple people fail to vote then I guess one is chosen at random, seems fairer). Basically, once the host PMs the bot "X is a replacement", the player X is put into a replacement list in the system, and is actually replaced when needed (and role PM is sent to him as well) | ||
gonzaw
Uruguay4911 Posts
I've finished designing the action resolution phase, and (apparently) it works fine now. It works only for the roles I'm restricting for this 1st iteration, so it won't work with bus drivers or 3rd party planar dragon and the likes. Use cases are done as well (they were documented here too). Role creation is done (in general terms, we can get into specifics in the design phase), as is setup/game creation. Chime in on OMGUS to discuss the final touches so we can start the next step. | ||
Nisani201
United States1400 Posts
On November 29 2012 10:32 gonzaw wrote: Okay I think I have everything under control now. I've finished designing the action resolution phase, and (apparently) it works fine now. It works only for the roles I'm restricting for this 1st iteration, so it won't work with bus drivers or 3rd party planar dragon and the likes. Use cases are done as well (they were documented here too). Role creation is done (in general terms, we can get into specifics in the design phase), as is setup/game creation. Chime in on OMGUS to discuss the final touches so we can start the next step. Can you show me what you have now (either on OMGUS or here)? There should be a clear place where we can see what the progress is. I was about to work on my own action resolution thing in Lua., but I'd like to see yours as well. EDIT: found the post you made where you talk about how it works, reading it now | ||
gonzaw
Uruguay4911 Posts
I.e vet/vig/cop/medic/RBer/Framer/etc (the ones I posted before) It won't work for other roles (or rather it's not planned to work), even if the whole "Create a Role" stuff I posted before would seem it does. Also yeah the main place where I post progress is OMGUS....but nobody reads nothing there so I have to post it here :/ | ||
Nisani201
United States1400 Posts
| ||
gonzaw
Uruguay4911 Posts
You just go to every thread that's red and read the updates. In the next phase we'll organize and tidy everything with models and nice drawings and stuff :3 EDIT: Also I didn't make each thread with the idea of updating them, I made them thinking about using them to discuss, and then update it with our conclusions... ...but yeah if I'm the only one posting it's just me updating so that didn't work >_> | ||
Promethelax
Canada7089 Posts
| ||
gonzaw
Uruguay4911 Posts
It's in the OP now. Here http://www.omgus.net/forum/viewforum.php?f=51 | ||
gonzaw
Uruguay4911 Posts
Imagine a certain scenario, X is a RBer, Y is a tracker, Z is a VT. Now X RBs Z, and Y tracks X. In this scenario Y would get a "X visited Z" report right? Good Now imagine X RBs Y, and Y tracks X. Does Y get a "X visited you" report? No, he'd get a "You were RBed" one since he was RBed, right? But how is this possible? Think of Action Resolution as "roles going out of their houses in a certain order to act", and think if roleblocks as "locking someone inside their house". In scenario 1, what happens? X goes out of his house and goes to Z's house, and locks Z inside. However, Y was already following X before that happened! What does this mean? It means that first of all, Y went outside his house, and went over to X's house. Then he saw X go out and followed him. He saw him go to Z's house, and then saw him go home. After that Y goes home and gets the "X visited Z" check. This makes sense right? Now let's go to scenario 2. Here, X goes out of his house and over to Y's house. X locks Y in and returns home. Y tries to leave his house but his door is locked, so he gets the "You were RBed!" message. Makes sense right? Absolutely not Tell me, generically, when does Y attempt to leave his house? After or before X leaves his? In scenario 1, Y tries to leave his house before X leaves his, that way he can track X. In scenario 2, Y tries to leave his house after X leaves his, because X already visited him to RB him. So the paradoxical question: How does Y know when to attempt to leave his house without knowing who X is visiting? If X visits Z, then Y attempts to leave his house before X left his If X visits him, then Y attempts to leave his house after X left his Thinking "realistically" this doesn't make sense and is a paradox. "But gonzaw! Why the hell does this matter? 'People leaving their houses' is just a conceptualization of what these roles would do in real life, but in the mafia game these are actions the host sees and can easily identify who visits who in the fly without following any specific order!" Yeah, the problem is that my system follows this "conceptualization" pattern, which is the one that makes sense. You know the Action Resolution order: first people get RBed, then bus drivers drive people, then framers frame people, then Cops get checks, then other stuff, and then KP goes through. This can easily be modeled with the "people leaving their house first" model, and is exactly what happens in my system. The problem is that if that model fails then so does the system....at least if it attempts to follow that model to a T. Yes I have an ugly way to fix that, but I wanted to know if there's "another" way to use this "model" to determine how RBers and trackers/watchers behave. Basically, if you think of trackers/watchers/Rbers as real people who "leave their house" at a certain time to perform their actions, how would you solve the tracker problem? Anything can be added to it to work, don't worry. Maybe you say something like "Every night, right before people go to sleep, the RBer gives his target a phone call telling him a story. This makes his target more sleepy than usual so he's guaranteed to wake up after the RBer.. After going to sleep an waking up at night, the RBer follows the usual pattern, while his target is guaranteed to be asleep when he goes to his house". That could easily be "translated" into a specific design of my system. The thing is yeah I just thought of that so I don't know if it would work (specially with other roles, for instance other RBers) so I dunno. EDIT: Another possibility is coming up with a "different" model that states how roles act. But yeah that's not very easy. | ||
Acrofales
Spain17834 Posts
Watcher/tracker are not actually watching or tracking. They are actually police dogs. They come in after the fact and sniff out who was at their target's house/where their target went. They were not actually there at the time of the deed, but could puzzle it together afterwards. Obviously if the polic dog was locked in his kennel, he couldn't get out there and do his job... | ||
gonzaw
Uruguay4911 Posts
I was thinking about doing something like: After every action is done, the tracker/watcher goes to his target's house and asks the guy "Who did you visit last night?" or "Who visited you last night?" But that wouldn't work since if that guy died by KP he wouldn't be able to ask him since he's dead, so I didn't consider that (basically making trackers/watchers the last guys in action resolution, even after KP. Rules of the game dictate you can't target a dead guy, so they wouldn't be able to watch/track him EDIT: Well....at least you can't do "normal" stuff on a dead guy.) However if what he tracks/watches is the house itself and the house itself has the "scent" of each visit somehow, that could work. The scent would disappear after that cycle is done, and a new scent would appear next cycle. Yes yes..that can indeed happen. I guess you were being sarcastic but thanks anyway lol | ||
Keirathi
United States4679 Posts
Just have it work the same as the regular system. If you are a tracker or a watcher, and you get roleblocked, you would still technically have a "result" but it just wouldn't be relayed to the player. When you send the player the PM, you check if he was roleblocked first, THEN if not roleblocked, send them their result. | ||
gonzaw
Uruguay4911 Posts
Doesn't seem too consistent though, since "technically" the guy shouldn't receive a check. If you use a "Get all Results" function you'd receive "Tracker got that X visited Y" even if the tracker was RBed, which "technically" didn't happen so the host will have erroneous information. There are ways to get over that....but yeah the above thing (the "police dog") thing works better I think. Like i said before I prefer to overthink it rather than overcomplicate it (in the implementation) and/or making the implementation "untidy" and/or making the system inconsistent/have erroneous data. | ||
gonzaw
Uruguay4911 Posts
Also are you experienced in working with UML? If you are then please head over to OMGUS if you can. There's not much activity in the respective forum of OMGUS (i.e I'm just talking to myself again), and I need to discuss this stuff since it's not quite as easy as I thought, and could make our lives easier in the future if we model it right (and I'm no expert in that). You may need to read all the walls of text about the requirements first though (so you know what you are supposed to model).....so yeah you may need some patience :/ | ||
| ||
![]() StarCraft 2 StarCraft: Brood War Dota 2 League of Legends Counter-Strike Super Smash Bros Heroes of the Storm Other Games Organizations Other Games StarCraft 2 Other Games StarCraft 2 StarCraft: Brood War
StarCraft 2 • musti20045 StarCraft: Brood War![]() • davetesta26 • Kozan • IndyKCrew ![]() • sooper7s • Migwel ![]() • AfreecaTV YouTube • Laughngamez YouTube • intothetv ![]() • LaughNgamezSOOP Dota 2 League of Legends Other Games |
Replay Cast
OSC
PiG Sty Festival
Clem vs Bunny
Solar vs Zoun
Replay Cast
BSL Nation Wars 2
Korean StarCraft League
PiG Sty Festival
herO vs Rogue
ByuN vs SKillous
SC Evo Complete
[BSL 2025] Weekly
Replay Cast
[ Show More ] SOOP Global
ByuN vs Zoun
Rogue vs Bunny
PiG Sty Festival
MaxPax vs Classic
Dark vs Maru
Sparkling Tuna Cup
BSL Nation Wars 2
The PondCast
|
|