Max Calculator Thread

[Hedgehog]

I received several PMs regarding this after I posted about it in the 200m in all skills thread. This thread will serve as a headquarters for the research and mathematics involved. Remember, this is a huge project, we'll need several thousand data points.

 

Step 1: Obtaining XP rates

 

We need xp rates for every level in every skill. To make this a little easier, I've divided them into a few categories.

 

Level-dependent skills

1. Slayer (and other combats)
   
2. Runecrafting
   
3. Dungeoneering
   
4. Hunter
   
5. Mining
   
6. Fishing
   
7. Woodcutting
   

These skills increase in xp per hour with each level (or every other level) independently of the method of skilling. It is important to know the xp/hour of each of these skills at every level in order to create an accurate calculation.

 

We can currently calculate the xp/hour of mining from 41-60ish and 80-99.

 

Data will be gathered largely through accuracy research. Let me do the math once we get the accuracy. We'll have to find a good way to find the xp/hour of dungeoneering because of how skill-based it is.

 

Method-dependent skills

1. Prayer
   
2. Construction
   
3. Agility
   
4. Herblore
   
5. Thieving
   
6. Crafting
   
7. Fletching
   
8. Smithing
   
9. Cooking
   
10. Firemaking
    
11. Farming
    
12. Summoning
    

We can calculate the time spent doing all of these skills right now (with the exception of thieving before pyramid plunder becomes efficient). The most important thing that we need is how each skill is trained at every level. We'll then calculate xp/hour and add it to the max calc.

 

Skills that benefit other skills

1. Magic (Summoning/Mining & Smithing/Slayer)
   
2. Other combats (Slayer/Summoning)
   
3. Dungeoneering (Woodcutting/Fishing/Firemaking/Combat)
   
4. Fishing (Agility & Strength when using barbarian fishing)
   
5. Woodcutting (Firemaking when using adze)
   

It will be necessary to calculate the total time saved of all skills that train multiple skills. For example, C2 fishing might save 1 hour of fishing, .8 hours of woodcutting, .6 hours of fishing, and .3 hours of dungeoneering per hour of C2ing. If 100 hours of C2ing is necessary to max, 80 hours of woodcutting would be subtracted from the woodcutting count, etc.

 

These skills will be calculated last.

 

Step 2: Creating the calculator

 

The coding of this isn't hard, at all, so I won't need much help.

 

I think the best implementation is to do a javascript based calculator on the web. I'd like to create functions to calculate xp rates, so that we don't have to store a ton of stuff in RAM.

 

You can calculate the hours to 99 using the function below:

 

 

Then it's basically just a matter of adding the hours per skill and then subtracting the overlapping hours from skills that train multiple skills.

 

What we need to do now:

 

Right now, we need hit/miss data for the first category of skills.Basically, just cut logs (or whatever it is for your skill of choice) for about an hour and record how many times you hit and how many times you miss. You'll need to keep track of ticks for some skills, a cannon helps with this.

 

Please do not discuss anything related to coding yet, it's just a distraction at this point. And let's focus on the hard to research skills first. I don't want people to run calculations on the easier skills then have everyone quit when it's time to do the hard ones.

[Quyneax]

You also need to determine the optimal order to max. For example, getting 92 firemaking before starting on woodcutting, or getting chaotics before maxing combat.

 

Another thing to take into account is obtaining xp-boosting sets and doing quests. For example, Smoking Kills before starting slayer, but also the 400k xp from WGS.

 

A third thing to take into account is people who are starting to use the max calculator, but who already have stats, which make certain orders or certain methods inefficient. For example people with 99 firemaking and 80 woodcutting, or people with maxed combat and 90 slayer, or people with almost all 99 just missing farming and construction or similar. This means the optimal order must be calculated live for each user.

 

Ideally, you would also take into account completionist cape requirements and side-goals such as hybrid, but that is a long way off.

 

Above reasons are basically why there is no maxing calculator yet. Simply combining 25 skill calculators doesn't come close to determining efficient ways to max at lower levels. On the other hand, calculating time to max starting at 80 base with quest cape is barely interesting, as most skills have narrowed down a lot by then.

 

By comparison, hit and miss data are relatively unimportant. If you can get the optimalization code working, you can plug in a number of approximate xp rates even without knowing the average ticks to level or xp per tick or whatever. Hit and miss data only really help to calculate lower-levelled xp rates, but that is less important than a working calculator.

[Platinum_Myr]

Doesn't thieving at pyramid plunder also involve success rate based on level? or is that factor minimal?

[Quyneax]

Yes. Cooking too, if you count burn rate, and agility too, counting failure rate, and crafting if you were doing gem cutting at really really low levels. I'm pretty sure Ape Atoll is good from like 60-75 even though you fail before 75.

[Hedgehog]

Pyramid plunder is level independent, except for the rooms that you can access.

We'll need to find burn rates and fail rates for cooking and agility, I don't think it's necessary for crafting.

You won't be able to be less than 0 hours from 99, so it doesn't matter if a skill is maxed.

I think the best way to handle quests is to only include them if they're the best way to train a skill (for example, waterfall for attack xp when you start).

In theory, a calc for all 25 skills is enough to create a max calculator, provided that you account for xp gained in other skills while training.

Completionist requirements should be ignored because it's a max calculator, not a comp calculator.

 

I don't want to invest time into creating a shell if it's not going to be as accurate as possible, so I think we should start with research.

[Quyneax]

Not true about points 3 and 5, and point 6 is subjective. I think a comp calculator is just as useful as a max calculator, so if you're making one, you might as well make the other. There is no reason to limit a project like this because you want people to do skilling research :P.

 

It matters whether a skill is maxed because you do not want to mine+superheat if you are 99 mining or smithing, but it's best if you are not 99 in both. There are a lot of multi-skill methods that stop being efficient if you max one or more of them. Another thing is that getting the gold mining suit stops being efficient after a certain time (not sure, maybe around level 85?) but it is worth getting for mining if you are sub 80 mining. I don't know if it's worth getting the suit if you are superheating.

 

And no, creating a calc for 25 skills won't do, exactly because of the order. You can't say 'okay, now get 99 all skills using the following methods: ..." and be done with it, you need to provide an order. A very simple example: for 99 firemaking, do you a) grind out all of 99 with pyrelord or b) use adze until 99 woodcut, then grind with pyrelord or c) use Book of Char, Char's Cave, Circus? If b, do you have 92 firemaking yet? How much woodcutting xp do you have left? If you're 98 woodcutting, you'll need to do more pyrelord and less adze than if you were 80 woodcutting. Same thing with C2 woodcutting + fishing (best to not max one before the other), but c2 woodcut will also reduce the fm xp you get from adze, if any. And farming needs to be added in between as well, but depending on how much time you have left to max, it may be more efficient to do cheaper trees because you have time to spare anyway. If you're going for comp, major xp gainers like 120 dg and Livid need to be taken into account, and the efficiency of for example Remote Farming and elite tasks and their rewards.

 

Plenty of good methods are also quest-locked (adze for one, but also PP and barraging), a bunch of slayer tasks are (dark beasts and ice strykes mainly), some areas or methods are skill-limited or miniquest-limited (barbarian course, certain parts of dungeons, Morytania, MTK/pengs/tears/troll invasion), shortcuts can speed up skills, and prayer can speed up or slow down depending on the availability of an altar in your poh.

 

The hours till max is not relevant, anyone could say '10k hours' and you'd be able to max in exactly 10k hours using some combination of methods. You need to determine an efficient way to max, and ideally the most efficient way to max. To be useful for many players, you would need to determine the best way to max starting at any combination of levels.

[Hedgehog]

Alright, so you'd like it to report what methods you should do? That's certainly possible too.

 

It would have to be able to generate the fastest route to max, then calculate the time it takes. We can also add a comp option to allow comp reqs and rewards to be factored in.

 

I still don't want to bother with any algorithms until we get the data we need, otherwise it's just a waste of my time.

[Quyneax]

Alright, so you'd like it to report what methods you should do? That's certainly possible too.

Wait what, you were intending to make a max calc that doesn't tell you what methods were used to calculate the time to max? What good is that :P?

 

Anyway, there is basically no incentive to do research on hit/miss when you can plug in experimental xp rates and approximate for lower levels from 10 minute samples. I don't need to know whether PP would theoretically average 156.728 xp/h at level 81, I need to know that I can get 150k at 81 and about 1k extra per level up to 90 (or whatever the figures are). Getting the time to max to within 5-10% of the real value is good enough (I think that most people are between 1-10% slower than ideal rates anyway, so a tick-perfect calc would be off as well), after that you can begin to refine stuff with hit/miss data. You say it's a waste of your time to code this without data, but it's a waste of everybody's time (a lot more, too) if the research gets done and no calc results. It's like building the supercollider before inventing the smartphone.

 

It's really great to see someone work on this, but remember that Tip.it crew has tried before and didn't get the optimization done. It's not as easy as it looks, with daily challenges and D&Ds and effigies and quests and whatnot. Making a calculator which uses perfect methods but doesn't take into account any of these is 100% theoretical and won't see much use.

 

Anyway for skills like Dungeoneering it may be best to provide a slider bar going from 'solo only, 12 min meds' to '5:5 only, 10 min larges' so people can pick what suits them, because there's just too many options to really pick one. There are also quite a few people who pick slower methods on purpose, so multiple options in general would be useful, but for dg (maybe slay too) it can save you a lot of trouble :P.

 

(lot of criticism from someone who hasn't done any work on any maxing calc, but you see why I won't touch it maybe :P)

[Hedgehog]

Well, I'm only doing this so that we can either get a max calc or end all of the annoying discussion about it :P

 

I still think research should come before algorithms. The algorithm isn't the hard part. The hard part is having an accurate calculator. What good is it to have a calculator that's 100+ hours off?

 

If you want to make a calc that tells you the fastest way to max using approximations, sure, but I can tell anyone that in about 5 minutes.

[Quyneax]

Nice thing about calcs is that you (e.g. you, Meredith, not you, generic user) don't have to be online :P.

 

Anyway about research, regardless of preferences, I can almost guarantee you will not get a lot of researchers if you have nothing working to show yet. Even if it's something 100+ hours off (which by the way is still pretty good, considering max is about 2000 hours/80 days or in that area). Proof of concept is very encouraging.

 

Really the time to max isn't so interesting (people change their plans all the time and get messed up by rl, you'd have to recalculate every other day), and the point where you can tell people in 5 mins is when they're about 90+ anyway. The interesting part is what you should do when you're low-mid level. Overloads or turmoil first? Rush for drygores or train all styles evenly? Go for torstol farming, but do I do calquats, and what about fruit trees? And most importantly: which (mini)quests do I do, what things can I ignore, when should I do WGS and what do I use 400k xp on? What do I do with Ports, dailies/weeklies/monthlies, how do I use Tears as efficiently as possible, what challenges do I do, what tasks (slayer and others), do I bother with ee? All these things are pretty tough to answer with a calculation. Usually we can ballpark it with intuition and experience, but that's hard to program. The smallest details, like going for EE instead of just using t70-75-80 2h (ee is a few % extra melee dps at 70-89 melees?) are probably too small to implement anyway, but that's the kind of thing you need to calculate to get to the accuracy level you want to go for.

 

Remember significance: 3.1 * 1.578904 (measurements) is still only 4.9. If your algorithm can't get close to perfect, it won't matter how accurate your skilling xp/h models are.

[Mojo_Monkey]

Right now, we need hit/miss data for the first category of skills.Basically, just cut logs (or whatever it is for your skill of choice) for about an hour and record how many times you hit and how many times you miss. You'll need to keep track of ticks for some skills, a cannon helps with this.

 

Need a bit more info on how this works, maybe direct me to a video?

 

I agree with Q that proof of concept will be a good motivator.

[jayc3399]

All the problems Q pointed out are why the spreadsheet I use relies on you picking a certain xp rate for each skill. It may not be exactly accurate but when you chose to do a faster more expensive method it just shows a reduction in time required. Maybe you level up and a new method comes available. You select that and your hours remaining plummets. Like I said not perfect and the first iteration didn't have selection. You just did what I deemed what was best and it was still nice for people to see as I did the most common methods.

 

 

[Quasar]

You should be warbanding mining in any case, and possibly smithing as well...

[Epilef]

IMO, firstly we need to gather information about every skill individually and after that we should make "single-skill-calculators".

Then add all these "single-skill-calculators" together, making a "max-calculator". Obviously the number of hours will be a few hundred hours off, but at last we will have something to work on to create an optimized calculator.

[Gemeos2]

Remember to not assume PP from 21 to 99. Blackjacking is the best method until level 91.

[Thai_tong]

Sadly these calculators are bound to come under criticism about their accuracy no matter how accurate you make them. I am actually most interested in how to simplify the summation.

Edited April 26, 2013 by Thai_tong

[Superheat]

Interesting read, I am certainly willing to test out some methods, as for instance I don't mind doing 1-99 Runecrafting again as fast as possible. As Quyneax stated though, it would be nice to see something already as I don't want the time to go to waste.

 

If you ask me, we should first find out what skills to do in which order as precisely as possible. As well as determine which quests should be done and in which phase of the game (e.g. Waterfall quest before doing anything else? DT as fast as possible or waiting until your levels are high enough to rush the quest?).

 

If we can get most of this figured out I can/ would be willing to take a few skills from 1-99 again as efficient (fast) as possible. Which would provide you with accurate data.

[Miss Lioness]

Interesting read, I am certainly willing to test out some methods, as for instance I don't mind doing 1-99 Runecrafting again as fast as possible. As Quyneax stated though, it would be nice to see something already as I don't want the time to go to waste.

 

If you ask me, we should first find out what skills to do in which order as precisely as possible. As well as determine which quests should be done and in which phase of the game (e.g. Waterfall quest before doing anything else? DT as fast as possible or waiting until your levels are high enough to rush the quest?).

 

If we can get most of this figured out I can/ would be willing to take a few skills from 1-99 again as efficient (fast) as possible. Which would provide you with accurate data.

 

Yeah I agree, also I would leave the combat stuff for last due to the rapid changes past few months et all..

 

Besides, Runecrafting for instance can be done at almost any time really. Prayer should be done as soon as you got access to a gilded altar of your own as this is faster. A tokkul-zo isn't required, but could help a tid bit. It will save only an hour at most over Ring of Duelling tho.

 

One should start dungeoneering once he/she got 80s combat and then go straight for chaotics. Just Staff and Maul.

 

Agility could be done at any time too as the courses barely require any quests and it will help with shortcuts and whatnot..

 

For the rest the first aim would be doing to quests till the point that it unlocks transport abilities. That means FTP2 till a certain point + The Grand Tree + Love Story for tablets. Also Heroes would be helpful to recharge glories but that can be fixed by asking someone with a Geyser Titan.

 

Any other general ideas to kick off with leveling?

 

EDIT: I am now skimming through quest lists to see what quests would be good to start with. Will post a list off the basic quests that will come in handy for sure!

[Miss Lioness]

Here is a list of quests that I would think would definitely be helpful to start off the game.

Not in order, just posting to see if anyone as any ideas or suggestions. Will try to change the list later after ideas/suggestions to be more optimal.

 

Animal Magnetism

Biohazard -> requires Plaque City

Buyers and Cellars - For later and quick access and Vial of Stench niche use.

Cold War

Death Plateau

Dwarf Cannon

Enlightened Journey - Transport + quite a bit of exp

Fight Arena (quest) - Lots of Attack exp

Firemaker's Curse, The - Albeit high reqs good quest to do asap

Fishing Contest - Shorcut

Fremennik Trials, The - Lots of exp rewards

Gertrude's Cat - Needed for Pyramid Plunder later.

Grand Tree, The - Long quest but unlocks lots of transportation namely spirit tree and gliders

Grim Tales - Short quest requires Witch's House also short but gives lots of exp

Hand in the Sand, The - Good niche income

Icthlarin's Little Helper - Needed to unlock Pyramid Plunder

Lost City - Not a too long of a quest, access to Zanaris required for Fairy Rings

Monk's Friend - 8k wc exp to kick of requiring nothing

Plague City - Long quest but unlocks some usefulness: ardy tele + west ardy

Sea Slug - Short quest gives a bit if fish exp | not needed at all

Tears of Guthix - Exp D&D

Temple of Ikov - Till you can get the boots, that is very very early in the quest

Tree Gnome Village - Complete spirit tree system

Wolf

The World Wakes - Lots of exp reward for it not requiring much besides gear!

[Thai_tong]

How about we just assume they have done all the quests so that an individual doesn't have to state every quest they haven't done... You're getting caught up on all the small things.

[Miss Lioness]

How about we just assume they have done all the quests so that an individual doesn't have to state every quest they haven't done... You're getting caught up on all the small things.

 

The point for the calculator is to include all the efficient things on the way to be maxed as soon as possible. Then one has to consider quests as they often unlock things that makes the road to maxing either a lot faster or a lot more convenient. The quest list I posted certainly adds the either of these aspects some even both.

 

For example: Dwarf Cannon quest. This allows one to use the Cannon which, once addressed, would help massively for slayer and saves way more time then the time it takes to complete the quest.

 

Quests are overall meant to support the player overall. Certainly balance must be sought for the trade off point.

[Thai_tong]

Then one has to consider quests as they often unlock things that makes the road to maxing either a lot faster or a lot more convenient.

One doesn't have to consider them if neglecting them has a small effect. Getting caught up on small details just leads to a burn out trying to account for everything and neglecting bigger parts of such a project.

[Miss Lioness]

Then one has to consider quests as they often unlock things that makes the road to maxing either a lot faster or a lot more convenient.

One doesn't have to consider them if neglecting them has a small effect. Getting caught up on small details just leads to a burn out trying to account for everything and neglecting bigger parts of such a project.

 

True, there are some quests I listed that isn't needed....

[Thai_tong]

Then one has to consider quests as they often unlock things that makes the road to maxing either a lot faster or a lot more convenient.

One doesn't have to consider them if neglecting them has a small effect. Getting caught up on small details just leads to a burn out trying to account for everything and neglecting bigger parts of such a project.

 

True, there are some quests I listed that isn't needed....

Would you read what I said. I didnt say that doing those quests wasn't efficient I said that accounting for them is a waste of effort.

[Quyneax]

Not accounting for them is about as inaccurate as not accounting for hit/miss data (just using 5-10h test for xp/h), so it's probably best to ignore that too for now :P.