How difficult would this be?

Ok, so with the upcoming update approaching us, a hot and almost overposted topic has been staking. What to do, what stats to use, will this be good? should I make a staker? ect...........

I'm not much for programming, but runescape's fighting system works off of bonus's that we can see, defense makes more attacks miss, attack makes more attacks hit, strength just controls the damage range that the system pulls from with a successful attack. With all of the difficult variables out of the way such as eating or healing in any way, with just the basics, attack bonus, strength bonus, attack lvl, strength lvl, and defense bonus, and defense level, is it possible to build a simple runescape fight simulator?

Basicly, enter the attack/strength/defense levels and bonuses for both opponents, choose how many fights you want to simulate, and have the program run it your chosen amount of times with a setting that deems the account that reaches 0 hp the loser and the other the winner. It would count wins for each setup, and tell you the number, so if you did 100 runs, it would run it and output like 40 wins for build 1, 60 wins for build 2.

It would be buggy and inaccurate at first, and the rs system of damage and attack would be difficult to pinpoint in the beginning, but if a few controls were run in game to get some base values like how many times a certain attack bonus hits against a certain defense bonus ect....to create a basic spread of effectivness based on the 3 stats.

I hope this doesn't sound too confusing, but if someone with some programming experience could reply and just lemme know how difficult a program like this would be to construct, and of course if you are interested, I would love to do some in game stuff to do what I mentioned above about establishing some base line figures to start with.

 

Impossible.

Unless you knew how jagex calculates the hit/miss ratio along with how much damage an attack does.

 

I mean we have 3 variables here, strength, attack, and defense.

SO, if you compiled a list of hits based on these 3 at increments of lets say 10, so for strength we have 1, 10, 20.......99 thats 11 tiers, and with 3 stats, thats 33 tiers to think about.

so you could compile graphs as such,

GRAPH 1 (DEFENSE 1)

ATTACK LVL​

STRENGTH LVL___1_______10________20_______30________40_______50_______60_______70_______80_______90_______99
__________10
__________20
__________30
__________40
__________50
__________60
__________70
__________80
__________90
__________99

The above graph would be the possibilities based on strength and attack on 1 defense, then similiar graphs would be obtained for higher defense levels, it would be a working progress and a huge project but I think it would be around forever and would benefit everyone a helluva lot. For starters each section, (i.e. 1 attack and 1 strength vs. 1 defense) could be taken on by one person. You simply get a 1 att, 1 strength account, and vs it towards 1 defense account with no weapons at first, just base stats, and do lets just pick an easy figure, so 100 hits, record what each hit was, whether it be a 2 or a 0. So after that is done, do other sections as well. Once each section is done, the main program could retain this data. So if the input into the program was 1 attack 1 strength 1 defense, against 5 attack, 1 strength 1 defense, the 2nd accounts options would be more desirable. It would work off of random number generators, so off of the recorded 100 hits, the program would randomly choose an option from the 100 choices based on the input stats. It would then subtract each choice it made from the opponents hp, and the opponent would do this as well. When an opponent reached 0 hp, it would do it again until it had calculated/simulated the specified number of fights, then output the results.

For starters, it would only be able to do no weapons, so the hit speed would be constant, it would basically just hit at the same time (as in it would pick a random hit from the compiled list at the same time increment as the opponent), but later when more choices could be implemented into the program, it could have other weapon speeds programmed in, so it would know to have certain weapons choose from the damage pools faster than others based on there relative speed.

I know this is probably confusing, so sorry for anyone reading this, just a kool Idea I had and thought with the right application it would be possible, quite a bit of work but as I said, hopefully the community could contribute by running these test pools. Would be mighty beastly down the road if it got integrated to use multiple weapons, maybe even implementing switching like starting off with dds specs and then going to whip, like alot of duels go.

 

yes as the above said this is not hard to make, the only problem is you need the data from jagex that they use.

if you can those, you any programmer can make this without much trouble.

an alternative is to get a less accurate version, where you actually do it manually with trial and error ingame and record your findings etc, but this wont be 100% accurate for obvious reasons.

 

I believe if enough data is collected, a decently accurate system could be created, yes it wouldn't be accurate down to the last kill, but even in an inaccurate system, say you tell it to run 1000 matches and 800 are won by one build and only 200 by the other, I would say that that is still conclusive to which one is better or would win more in that situation.