Project C.A.M.P.

[Corvus]

Update: Due to a lack of submissions, we have made the decision to cancel Project C.A.M.P. We do thank everyone who took time out of there schedule to submit data.

 

Original Message:

Hello Tip.it,

 

All of your friends here on the Tip.it Crew would like to announce the start of Project C.A.M.P. (Calculating Accuracy Modifier Percentages), an attempt to figure out how accuracy within Runescape works, which, if with your help succeeds, will allow us to provide you with new tools to help you in-game. We need your help with this project because of the massive scope of the research that is needed. Currently we are planning 3 seperate batches of research on Melee combat that will allow for us to create a calculator that will provide a rough estimate of your accuracy during combat. Currently we are looking for any players who are willing to spend some time fighting Cockroach soldiers to gather data for us. Players of any Attack Level are welcome even if someone else of your attack level has already replied. Thanks to everyone who is willing to take some time to help us with the First part of our research (instructions are below).

 

From

The Tip.it Crew

 

 

How to help:

Setup: Unequip weapon and switch to the defensive style of unarmed. Remove any other items that modify the Crush attack stat and anything you do not wish to lose (3 items recommended). Must have access to the Stronghold of player safety (back entrance suggested).

 

Step 1: Enter the Stronghold of Player Safety.

Step 2: Combat the Cockroach soldiers inside and count how many times you successfully deal damage and how many times you deal 0 damage. It is important that you remain as accurate as possible, though small errors are acceptable. It is up to you to decide how many attacks you wish to count, but we ask that you record a minimum of 100 attacks. However, More than 100 attacks is always welcome as it will improve the accuracy of the data.

Step 3: Report your attack level, data, and the mystery phrase Pineapple here : Tip.it Project C.A.M.P. Submission Form

Step 4: Keep your eyes open for Part 2 of this research project.

 

P.S. We cannot promise that we will be able to provide any new tools at the end of this project because we may not succeed in determining the formula.

[Corvus]

Frequently Asked Questions

 

What equipment do you recommend?

I would personally recommend any 3 pieces of standard armor and no weapon, as this is the easiest way to get a defensive crush weapon and no crush attack bonus, though make sure you check the equipment menu to make sure you dont have a crush attack modifier. If you wish you may also use food which will extend the time that you can spend collecting data.

 

What weapons have defensive crush attack types?

Unarmed, Maces, and Warhammers all feature Block attack types which meet the defensive crush attack requirement, however Unarmed is the easiest to use as it provides no crush attack bonus

 

How can we keep track of the numbers of hits and missed attacks?

The easiest way is to keep a tally of each individual attack on either a sheet of paper or text editing software.

 

Why do we need to enter the Mystery phrase when we submit data?

The purpose of the Mystery phrase is to make sure that everyone submitting data has read the instructions on data collection, to help improve the accuracy of the data.

 

What data do you collect when we submit the form?

When you submit data we ask for your Attack Level, The number of hits, the number of misses, and the mystery phrase. Other than this data that we ask you for we only receive data about when a submission was made.

 

What should I do if I realise that I made a mistake after I have made a submission?

if it is a reletively minor counting error it should be fine, but if it is a rather major problem please post here or private message me (wyvren2000) with any of the following information you can provide: Date of submission, time of submission(doesn't need to be exact) and timezone, Attack Level, Hits, and Misses.

 

Why Cockroach Soldiers?

Cockroach Soldiers were selected at the target of this research because they are F2P, easily accessable, and are reasonably high in level and defence.

[Cowman_133]

This project is now open for user contributions. I'd like to thank you in advance for your assistance!

 

DK

[Gradeskip]

First post? :D

I'll gladly help!

[Warrior_B666]

I'll see if I can find some time to help out :smile:

 

At first I thought this project was related to something like camping at abyssal demons :lol:

[x_bow80]

Havent done much to help tip.it lately so ill pitch in when i can :)

[Jard_Y_Dooku]

Are you going to make the data available publicly? You're using Google Spreadsheets for storage so there's no excuses on your part about bandwidth or anything like that. ;) I hope this doesn't lead to another secrecy situation (like the Combat Calculator) where Tip.It finds a roughly accurate formula and then hides it from everyone which will stunt improvement.

 

Also, have you considered using the turmoil prayer to attempt to identify the attack, strength and defence levels of monsters? It will be able to tell you their attack, strength, and defence within a 10 level range - n to n + 9 where n is divisible by 10. This could actually allow us to find the combat stats, with a rough margin of error, of a lot of monsters. It would be very interesting to see what results we get. For example, consider a man and a chicken's stats. With all stats 1 aside from HP, their combat levels come out accurately (although the fact that chicken never hits is a bit fascinating). I think this also proves that the minimum maximum hit is 15 (according to monster examine) (assuming a strength level of 1) as they would appear to have no strength bonus. This could also help determine their strength levels more accurately. Something to think about...

[TheAncient]

Couldn't you just test the dual dice roll theory? (each player rolls a value based on the strength formula, except with defense and attack stats/values; if the attacker has the higher value the attack hits)

[etpelletier]

Already sent something :) 198 hits in total

I wish you good luck with this project, it sounds interestingly fun!! :)

 

If you need more help, just ask! :) One of my best friends met via RS is a part of the crew, and it's something easy to do for me since I'm now f2p wearing p2p only equipment.. no need to bother finding a gear :P

 

Etp <3:

[Napalm]

Will most gladly contribute tomorrow, as I'm too tired to do so today.

 

Best of luck with the project!

[Jard_Y_Dooku]

As a follow-up to my previous post perhaps someone might find this interesting...

https://spreadsheets...uthkey=CP7Okt8O

[Kata_Phfract__the_slayer]

i shal add to this.

i am thinking of ways to recording the hits easyer, if the ticker item does not interup combat it could be used to count times hit (0 or not) and just mark damageing hits on paper.

(edit: ticker does interups combat but you punch so slowly that you can quite easily fit it in)

to make trips last longer i guess rapid heal prayer + regen brace would be handy. but the brace adds +8 to crush, you could compensate that by equiping flipers + Armadyl plateskirt/Spined helm, or berzerker necklace + unholy simbol/gauntlets/"bronze" RfD gloves.

ect ect

 

or you could just take food lolz.

 

edit: realised i could just take 2 tickers and have 1 for miss, and one for hits :wall:

[SwreeTak]

I think I will contribute to this. I will go kill some cockroaches later today :)

[Markup]

CAPTCHA or someone will ruin it

[Tecmaster]

I'll start killing later today, a bit tired atm.

[Corvus]

Jard, at this point in time we will not be allowing unrestricted 3rd party access to the data we collect for this project. The Tip.it staff will re-evaluate this concern at a later date and a final decision will be made at that time. If you or anyone else wishes to discuss this matter further contact @Cowman_133. It is not necessary to contact me on this issue.

 

When the time does come for planning the phase of Project C.A.M.P. dealing with monster stats we will make sure to consider the Turmoil prayer as an option, however at this time we are working on establishing an initial result to allow us to use this data.

 

Sseli, we will be considering that as a possibility when we analyse the data.

 

Expall, due to the method of data collection we are using, Captcha is not able to be used on this, however we do have a multi layer method in place to prevent improper use of this form.

[AlexRose]

Well hold on, how do you know the defensive value of the Cockcroach Soldier, and how said defensive value would come into play? If the calculation regarding defense isn't a completely linear relationship with your attack level, then your results could become skewed.

[CookMePlox]

Hmm, I was going to do something similar to this. However you forget that sometimes you can still "hit" even if you hit a 0. You'll need to multiply the number of 0 hits by one less than their maximum hit divided by their maximum hit. I would think it would be easier to do this against players, where the defensive stats are known. I assume you're going to try to come up with a formula?

 

I did some more thinking on this, and came up with six variables on your accuracy:

 

Attack level

Attack bonus

Opponent's Defence level

Opponent's Defence bonus

Your Attack style

Your Opponent's Attack style

 

Those variables could be reduced down to four if we assume that Defence level/bonus has the same affect on accuracy as Attack level/bonus. (For example, the accuracy of an unarmed person with 99 attack hitting a person with 99 defence equals the accuracy of someone with 1 attack hitting someone with 1 defence).

 

For now, I think, it would be best to focus on how the Attack bonus affects accuracy. We know, of course, that the higher your attack bonus gets, the closer you get to 100% accuracy, but it never reaches 100%. Also, the lower your attack bonus gets (in the negative), the closer it gets to 0%, but never reaches 0%. I'm not sure what kind of function that would be, to be honest. Anyway, to test how the attack bonus affects accuracy, we'd probably want to test against players, instead of monsters, so the data would be useful in later tests on the other variables.

Starting unarmed, you'd test, say 2500 hits against someone else. You could do that in the Duel Arena or in a POH. Then you'd repeat the same test with, say, Flowers that have a -100 attack bonus. Then you'd do the same with another weapon that hopefully has +100 attack bonus (I can't think of one off the top of my head). You'd probably want to also test in between those numbers as well. I think that this variable would be the most useful to players.

 

Just a thought,

Cook

[Saradomin_Mage]

^ I assume that after the Constitution update, it's now impossible to hit for 0 damage, since in the past most zeroes were a result of a 0.1 to 0.9 roll being truncated.

[CookMePlox]

^ I assume that after the Constitution update, it's now impossible to hit for 0 damage, since in the past most zeroes were a result of a 0.1 to 0.9 roll being truncated.

 

There still is a possibility, however slim, of "hitting" a 0 and should be taken into account in a project where correct equations are tantamount. Now, if you roll between 0 and 0.1 (or a 0 and 1 in the terms of now), you still can "hit" 0.

[Corvus]

The issue is that while there is some chance of getting a miss from the damage half of attacking, it ends up that there is no effective way to separate the accuracy miss from failed damage (short of figuring out the exact way that damage is determined). Also yes it would provide a bit more defined data by using player vs. player combat, its just that in this circumstance we want a large number of replies, and it is far more likely to get those replies by having players fight a monster than having to track down someone with the correct stats who is willing to allow the player to kill them several times in order to get data.

[CookMePlox]

The issue is that while there is some chance of getting a miss from the damage half of attacking, it ends up that there is no effective way to separate the accuracy miss from failed damage (short of figuring out the exact way that damage is determined). Also yes it would provide a bit more defined data by using player vs. player combat, its just that in this circumstance we want a large number of replies, and it is far more likely to get those replies by having players fight a monster than having to track down someone with the correct stats who is willing to allow the player to kill them several times in order to get data.

 

If you recorded what their max hit was, because the hits are equally distributed across 0 to the max hit, you would multiply their number of 0s by (maxhit - 1)/maxhit.

[12pure34]

The issue is that while there is some chance of getting a miss from the damage half of attacking, it ends up that there is no effective way to separate the accuracy miss from failed damage (short of figuring out the exact way that damage is determined). Also yes it would provide a bit more defined data by using player vs. player combat, its just that in this circumstance we want a large number of replies, and it is far more likely to get those replies by having players fight a monster than having to track down someone with the correct stats who is willing to allow the player to kill them several times in order to get data.

 

If you recorded what their max hit was, because the hits are equally distributed across 0 to the max hit, you would multiply their number of 0s by (maxhit - 1)/maxhit.

 

That would mean you assume normal distribution among 0 - max hit. How sure can we be about that? P(Maxhit) = P(Hit=0) or not?

Before the constitution update, you couldn't be sure of that, since 451 would be equal to 450 (or 460 if it rounded upwards). Who says the damage isn't calculated in a way like 451.3 now? I don't think you can get the exact failed damage from maxhit-1 / maxhit due to this. It would give you a good estimate though.

[CookMePlox]

The issue is that while there is some chance of getting a miss from the damage half of attacking, it ends up that there is no effective way to separate the accuracy miss from failed damage (short of figuring out the exact way that damage is determined). Also yes it would provide a bit more defined data by using player vs. player combat, its just that in this circumstance we want a large number of replies, and it is far more likely to get those replies by having players fight a monster than having to track down someone with the correct stats who is willing to allow the player to kill them several times in order to get data.

 

If you recorded what their max hit was, because the hits are equally distributed across 0 to the max hit, you would multiply their number of 0s by (maxhit - 1)/maxhit.

 

That would mean you assume normal distribution among 0 - max hit. How sure can we be about that? P(Maxhit) = P(Hit=0) or not?

Before the constitution update, you couldn't be sure of that, since 451 would be equal to 450 (or 460 if it rounded upwards). Who says the damage isn't calculated in a way like 451.3 now? I don't think you can get the exact failed damage from maxhit-1 / maxhit due to this. It would give you a good estimate though.

 

I have done significant testing to show that the hits are uniformly (NOT normally) distributed from 0 to maxhit. However, P(Maxhit) will not equal P (Hit=0) because the max hit may be truncated, as you said. However, I would say P(Hit=0) = P(Maxhit-1). As you said, it is a good estimate, and improves upon the current system that the crew is using.

[Corvus]

Due to a lack of submissions, we have made the decision to cancel Project C.A.M.P. We do thank everyone who took time out of there schedule to submit data.