Jump to content
Search In
  • More options...
Find results that contain...
Find results in...


  • Content Count

  • Joined

  • Last visited

  1. I'm not at all programmer but I don't think you need to focus on specific AI for every creatures. Like you said it's a lot of work for a PvP game and it means MANY potential bugs to correct. For me the better system (and maybe the simpliest i don' t know) would be to have a large data base of creatures and plants (which is a huge artistic work i admiit it) with, like i said previously, specifics caracteristics. After this you can handle this concept more like a simulation than specifics AI. What I realy want to say is: "What is an ecosystem after all?" and "How can we use this notion to create this new world more realistic and realy living ?". In ecology you can theoretically take a biotope by its parameters because animal behavior is mostly created by reaction to this parameters. Instead of creating multiple AI you can apply simple natural, global laws of interest, what we call in biology a "gain/loss ratio". I know i'm going too far but if someone can read it with programming capacities i would like to know if it could be handled by our generation of servers/machines. I give you again an exemple of what i think would be the better and the most ambicious ecosystem in a MMO: How can we create realistic reactions for creatures ? And here we will take something more specific, a situation that you described BahamutKaiser: How a virtual animal could decide if he flees or attacks an enemy ? It's what we can call "ESS" in biology (Evolutionary stable strategy http://en.wikipedia.org/wiki/Evolutionarily_stable_strategy) this system can explain many things in the nature and after all it's just simple maths and probabilities, something that computers can handle. I'll take a simple case to illustrate: You are in Crowfall searching for strawberries in the forest (again this fruit sorry...). A tiger appears right in front of you, how can we make the tiger reacting like a real beast ? I) Parameters (again) I already did talk about this in my first post, but we can explore an another face of it. 1) Survivability The tiger, when the dev coded it in the data base, has been related to many numbers, for exemple: - Attack points - Armor - critical hit - regen - hp - speed .... We can create with these numbers a "survivability" parameter (it exists also for your character), your capacity of winning the fight is shown by this number. For this exemple let's consider random survivability numbers: You = 120 S Tiger = 90 S 2) Potential gain And i'd like to see a new key parameter (more or less complex): "gain". It's a very complicate and evolution-related thing in the real world but you can, in a game, create something simplier: Your level, resources in inventory, achievements... Could be synthetized into one variable (simpliest case) which represents the "gain" the tiger could win if it kills you. The issue is to imagine something simple to modelize the gain, how the monster could have a real interest of killing you ? Again i synthetise but you can imagine that in a fantasy world creatures can absorb a little bit of your power when they kill you, it's an idea of potential "gain". Again we choose random numbers: You = +60S (if we see gain like a + survivability bonus for the tiger) II) Let's calculate ! Ok so we have here two parameters, potential gain and survivability. With that we can imagine a simple formula to finaly know if the tiger will attack you or not. If Survivability potential is equal between two opponents the chances to win the fight are equal for both characters. S- = Death S+ = Power increases Sp = S of the player St = S of the tiger St/Sp Sp/St 90/120 = 0,75 120/90=1,33 coef bonus for the tiger coef bonus for the player Now let's create a basic gain/loss relation: If the tiger looses, he dies --> - 90S If he wins, he becomes way more powerful --> + 60S Possible formula of the potential Gain : Gp=(S- * (Sp/St)) + (S+ * (St/Sp)) Gp=(-90 * 1,33) + (+60 * 0,75) = -75 I forgot a last parameter: The condition: if the gain/loss ratio is < G then the tiger will engage the fight. G = -40 (he is quite agressiv = he can take risks, you can change G to change the agressiveness of the creature) -75<-40 --> the tiger wont take the risk to attack the player This system has many defaults, it's just to show that is not impossible to simulate conflict behavior between two mobs or 1 player vs 1mob. What i like with this kind of formulas is the fact that you can change every parameters to change the comportement of the creature : If you take G=200 you will have a mob who never wants to take a fight or only versus a very very easy opponent, it could be for herbivorous animals. If you take G=-200 you will have a very aggressiv mob, it could be for corrupted dangerous monsters. If you take G=0 you will have a normal predator.... The thing i don't know how to handle is the « cost » of the mob-life... It's difficult to attribute a number for it ^^ (i took the Survivability in this model but it could be something else) You can also estimate when a monster wants to attack you, and even when he will stop the fight. Because if you refresh during the fight the Gp value, the monster can decide to go away because his survivability is now too low → Gp<G again. In our situation we can see that the tiger would attack you only if you have a survivability above 90-95S or under, you can change every parameters you want. I think about these ideas for a long time now and i always wanted to see it in a MMO, it will not happen i know it but i had a good time sharing with you this sutff^^. I don't have a high level in maths/coding and not a pro in biology so if you see (and you will) huge defaults/inconsistencies in my propositions please share it ! Sorry for the long post, and my bad english as usual . Ps: I wanted to show that you don't need Specific AIs for each mob but i guess it's what i'm talking about here^^ Ps2: This model is an exemple for the simpliest case: 1v1, we can imagine more complexe parameters for more complexe situations.
  2. I think they eat it in peru.... Sorry guys
  3. Yes chat reglementation is a huge part of the immersion... (I did not pos tin the right subject so i edit in something random )
  4. The most important thing for me will be the "visibility" of enemies (and allies) since it's not pure RvR you have to deal with a lot of parameters to know if this guy will discuss or just throw a fire ball at you.... The worst would be the classic red names... I'm again speaking about fps (i have a long experience with PS2), because it's the only massiv strategic battles i've been to, but something like "i see an enemy --> i use a class spell--> me and my guild/commando see it for 30secs" would be the best. It's not realy useful to speak about commander skills if we don't know how will be presented the enemies in the GvG game Crowfall wants to be. Let's talk about alignment and alliances, it would be more productive i think at this state of the game ^^
  5. It has been done in few games like you said Raeshlavik, but why it never worked ? Maybe because these worlds did last longer than they should, an ecogolical (because yes it is) system, like that, has defaults it's obvious, and the more your system lasts the more it will be dizorganised and unbalanced. Crowfall can maybe limit this drift by these time-limited worlds. The interaction with the player is the hardest thing but you can maybe cheat a little: 1) Overpopulation of one mob in one zone --> the server detects it and start to kill random individuals like a natural plague 2) Subpopulation of one mob in one zone --> the server detects it and begins to "drop" this species in an another zone adapted --> or the world is time-limited and it's begin to be a survival: the server does nothing because you just had to be a better strategic player, such a hooligan this server. --> There are plenty of different mobs that provides the same resources (with maybe different properties after being crafted) so you can handle one extinction.
  6. You are right, if we have physics like described i think you can moove/destroy almost any items/voxel so it 'll be fun to be hable to moove and hide bodies (maybe for a res?).
  7. For this kind of thing i think planetside 2 is a good exemple. I'm speaking about abilities to create respawns/bonus in a squad/global orders and signals.
  8. Wakfu was a cool first try for introducing a real ecosystem in a MMO, the first beta version (2010 i think) was awesome. You had to manage every resources, guilds started to create laws for the gestion, blocus and monopolies were created. It was awesome but now this game has left all his ambicious features T_T. My dream would be to see something like this: 2 categories: Plants: 1) Life time-line based on the species (seed, growing, mature --> reproduction with seeds / spores, natural death) Because you need to create a dynamic if you want a simulated ecosystem to be interesting. 2) Ground and climate specificities (Temperature, altitude, wind, rain ?...) -Create specific temperatures, rainfalls, winds... generated by the position and the nature of the voxels (biomes as minecraft) -Create conditions for every plants, exemple: Optimal, possible and not possible (very basics categories) could be linked to seeds/resource production, time-line, % chances of successful planting or % chances of spawn at the creation. Every parameter could be linked to each other (if it's cold AND dry it would be worse than cold AND humid for exemple). Strawberry (we need sweetness too) Temperature: Optimal: [20;30] C° Possible: [10;20] or [30;40] C° Not possible: <10C° or >40C° Altitude: Optimal: < 1000m Possible: 1000m<h<1500m Not possible: >1500m ..... If you create simple datas like that for every plants and if you link it with your voxel world, ecosystem diversity will be created by itself. The more you have details and diversity of plants the better will be the complexity and the diversity of all the worlds/campains created. With this system you can also change these conditions during the campaign (to make the world harder in winter for exemple). 3) Create predation/competition - Trophic interaction with animals (the bunny eats the strawberries or the strawberries eat the bunny we never know in a heroic-fantasy world ^^) - Limited "resources in the ground", every plant should use some limited "resource points" around them. If there are not enought "points" for a new plant, it will not grow / the most vulnerable organism will disapear (vulnerability linked to the previous parameters of the place). Animals: It's almost the same data base but the population is limited by its need to eat others organisms (plants/animals/players). 1) Life time-line (egg/larva/immature, mature --> reproduction sexual / asexual, natural death) 2) Climate specificities + maybe essential plants. Again a simple data base which is linked with vegetals and voxel system. 3) predation/competition - Trophic interaction - If the animal/monster doesn't eat he dies... So if there are too many individuals for the food present in the biome population will decrease / the most vulnerable organism will disapear. 4) (bonus) Evolution ? Not a real darwin natural selection system but something like: "a population of tigers have been nearly decimated by 2 archers the last two days, the third day the two players will have to fight mutated tigers with... shells ? Or a new group defense strategy ? Or a specific Boss who dodges every arrow ? Could happen between species too. Bonus because it must be harder than the rest to create but it would be amazing. I think it would not be very complicated, i'm not a programmer (biologist in fact, you had perhaps realized that ^^) but it's only linked data base. After i don't have a clue about the performances needed to runs every interaction. This system works only if you have enought species of plants/animals created. But if you manage to have diversity with parameters and organisms this cocnept could create basic ecosystems (faaaaaar away from the complexity of our world^^) by itself with the world generation. I'll maybe develop this suggestion in a post later, it's something that could make a game unique and endless but most of the devs think this kind of feature doesn't bring players and money. Just think about all the interactions wiht the hearth of the game: PvP, it could approache and maybe surpass the complexity of EvE. Blocus, monopolies, rare resources, creating natural bareers of monsters/trees to stop the enemy charge, create agriculture and animal breeding for resources (mounts, familiars ?). Even why not an ecological war ^^: Guild A wants to destroy medecinal plants production of Guild B. Guild A captures some monsters who eat these plants and makes them reproduce. Guild A launchs a commando to release many monsters in Guild B garden. Guild A wins the siege because Guild B doesn't have potions anymore ! I dream i know but maybe a little something in this could be introduce one day in a game and i would be so happy^^ Sorry for the long post and the bad english (i'm french) ! Thanks for reading !
  • Create New...