The MATH of Archaeology

[The Pro]

So. I was thinking. Archy tells me to go to the nearest archaeology point, no matter what. But suppose your map looked like this.

There's no way that going to the bottom X will yield you the most fragments per hour. I've been thinking about this for days. I'm not sure if there even is a set strategy that will yield the most digsites per hour, and it might just be different for every situation. If there is, it has to consider the number of potential digsites the distance from whichever you go to, and the chances that one of these will spawn a digsite near yours. I have no idea: the closest one could always yield the best fragments per hour, and I could be wrong, I am really unsure.

I asked one of my teachers. He's the top paid math teacher (college professors don't count) in my state (I go to a private, rich white boy high school), and he's really smart.. He said he didn't know the answer without a computer program. Any ideas?

EDIT: I had a drawing. It failed. It showed 3 digsites near the top of the map, one at the botton, and the player was just SLIGHTLY closer to the one digsite at the bottom of the map than to the other three.

[Denzion]

I don't think it can be calculated reliably because the location of each new dig site is always random.

[shatterednebula]

What you're asking is a play on the Traveling Salesman Problem, except, like Denzion said, those locations would change. By the time you were to find an optimized solution for it, you would have had the time to level archeology 4 times over. Its best to just put on some music and grind a little at a time.

[Skelly]

Sounds like your teacher is being paid too much.. Its a pretty simple problem, and all other factors being random, going to the 3 sites that are close together will save you time.

If you factor in where your hearth is, where you can teleport to, and the areas where digsites spawn most often, the above can change

[The Pro]

What you're asking is a play on the Traveling Salesman Problem, except, like Denzion said, those locations would change. By the time you were to find an optimized solution for it, you would have had the time to level archeology 4 times over. Its best to just put on some music and grind a little at a time.

Yeah, he did mention something about that. Thanks for the reply.

Sounds like your teacher is being paid too much.. Its a pretty simple problem, and all other factors being random, going to the 3 sites that are close together will save you time.

If you factor in where your hearth is, where you can teleport to, and the areas where digsites spawn most often, the above can change

Its obvious that it's better to go to the three digsites.. I was searching for a general strategy to determine the best fragments per hour in any situation. My example was just to show that going to the nearest digsite is not always the best route.

[Jaerin]

So. I was thinking. Archy tells me to go to the nearest archaeology point, no matter what. But suppose your map looked like this.

There's no way that going to the bottom X will yield you the most fragments per hour. I've been thinking about this for days. I'm not sure if there even is a set strategy that will yield the most digsites per hour, and it might just be different for every situation. If there is, it has to consider the number of potential digsites the distance from whichever you go to, and the chances that one of these will spawn a digsite near yours. I have no idea: the closest one could always yield the best fragments per hour, and I could be wrong, I am really unsure.

I asked one of my teachers. He's the top paid math teacher (college professors don't count) in my state (I go to a private, rich white boy high school), and he's really smart.. He said he didn't know the answer without a computer program. Any ideas?

EDIT: I had a drawing. It failed. It showed 3 digsites near the top of the map, one at the botton, and the player was just SLIGHTLY closer to the one digsite at the bottom of the map than to the other three.

It's called the Travelling Salesman problem and there is no easy solution. This falls under one of the P = NP data sets that if answered would solve a lot of incredibly difficult math problems. The reason being is as you increase the number of potential destinations the number of possible paths go up exponentially. Now you can optimize it such that you eliminate poor options, but there is no way to currently know if you are or are not eliminating the most efficient solution by doing that.

In the case of WoW it should be pretty easy to do because you only have 4 points to decide between. The biggest obstacle is you have no way to predict where the new spot is going to be added. You would need to put in some kind of a weighting system that would not only figure in the current shortest path to the next point, but also a value of how clustered the points are.

So you should be able to take all points, measure the distance between that point and all the other points to determine their relative position to the others. Average those values and you will have a cluster value. This is how clustered other points are relative to that single point. You might also want to play with putting a maximum value on the distance between points to consider for this cluster value. This will prevent it from say skewing the cluster value when you have 2 points at one end of the continent and 2 points on the other end. Then you should be able to compare and weight the values of not only the distance to the next closest point, but also the best destinations after that point.

It is definitely not an easy problem to solve.

---------- Post added 2011-05-25 at 04:40 PM ----------

Actually probably a good start would be to do what I said and create cluster value for each digspot relative to all possible other digspots. This would give you an overall cluster value for a digsite.

You could then use that as a weight for which sites have the potential for more new sites near them. For example in the situation where I said there were 2 on one end and 2 on the other, it might be that one pair has more non-activated digspots. That could reasonably be used to predict if the next likely point will be near it or far away. If all 4 points are in the north the likelihood that the next point also being in the north is smaller because 4 sites are already active and we know that the next one can't be one of those.

The hard part is coming up with proper weighting for all these things.

[Skelly]

Like Jaerin said, you don't have enough information to get the optimal solution. Best practice is to go towards clusters (go to the three at the top in the original example). In the example of two at one end and two on the other.. Who cares? By knowing the optimal solution you'd save yourself a few seconds. But look, you're doing archy which is wasting a hell of a lot more than a few seconds

[MrSerious]

But look, you're doing archy which is wasting a hell of a lot more than a few seconds

^ this.

why bother? its not challenging, its not fun (flying back and forth repeating a mind numbingly simple task), its time consuming, its only good for doing something when you have absolutely nothing to do and thats the case people need another hobby than just wow.

[Cairm]

Don't think too much, its a secondary profession. Its self-explaned with : Go to what ultimatly will yeild you the more site/hours. In this case, the 3 site that are clustered.

Too many variables that are random are being used, and thus, like Jaerin said, its the Travelling Salesman issue where the amount of possibilities will go up exponentially.

---------- Post added 2011-05-25 at 09:59 PM ----------

^ this.

why bother? its not challenging, its not fun (flying back and forth repeating a mind numbingly simple task), its time consuming, its only good for doing something when you have absolutely nothing to do and thats the case people need another hobby than just wow.

I soo agree with you. Thank god.

[morpsend]

We are working on a time per solve project. Any information you come up with would be very helpful!


http://www.mmo-champion.com/threads/...time-per-solve.
http://us.battle.net/wow/en/forum/topic/2522464295
http://wowdigsite.com/2011/05/player...ime-per-solve/

[Jaerin]

^ this.

why bother? its not challenging, its not fun (flying back and forth repeating a mind numbingly simple task), its time consuming, its only good for doing something when you have absolutely nothing to do and thats the case people need another hobby than just wow.

Because people find enjoyment in the game other than what is just presented by Blizzard. There are the people who love theorycrafting, the strategists, the collectors, the achievement whores, pvpers...all kinds of people.

Some people just like to think about these problems and try and solve them. And as morpsend pointed out below above me it looks like others are already looking into it.

[meldora]

There is probably a calculable 'random average distance' ~ if there is a digsite with another digsite near it, where the distance is less than the random average distance, than that digsite would be better to go to than the digsite with the nearest digsite being further than the random average distance. How much better would depend on the difference between the RAD and the difference in distance between the two digsites. That's all I got

[Deleted]

Its obvious that it's better to go to the three digsites.. I was searching for a general strategy to determine the best fragments per hour in any situation. My example was just to show that going to the nearest digsite is not always the best route.

no. it is also possible to go to the south location, continue using that spawner until it spawns near the other 3 locations, then use hearthstone, port hyjal(kalimdor), undercity(horde), epl(dk), and do all 4 locations. if none of the aforementioned is possible, then it might be more efficient to instantly do those 3 areas first.

edit: also what meldora said above this^^ however you have to factor in the possibility to teleport around.

[Porcell]

I use a "herding" method. Lets say I have one up in hinterlands, one in badlands, and two down in STV. Starting in SW I'll go down to STV and clear those two. If any more pop in STV I'll clear those. Once everything in STV is down, look for the closest, like if there's one in Swamp of Sorrow or Blasted Lands. If any are in close flying distance, I'll go to them. I'll keep sweeping back and forth across these four southern zones until they are all clear and gone. At that point I will have "herded" the dig sites more to the north. If I'm hardcore digging, I will usually have my hearthstone set to Hinterlands, so I can just Hearth up there and start herding sites back to the south.

I can hearth to the north and I can teleport multiple times back to Stormwind (guild cloaks, TB tabard) to hit up the south and central zones.

Best optimizing your available teleportation is the only way to really speed things up.

[It's interesting to me that Chrome plugin thinks teleport and teleportation is not a word.]

[Kragragh]

I get your point in all this, and if you enjoy that kind of math, by all means knock yourself out. But, I think there's too much randomness to really be able to get good information.

I think you'd be better served in creating a "plan" to minimize your down time. i.e., make sure you have Hasty Hearth, if you're in EK set hearth here, make sure you have teleport cloaks, if you're engineer then use your wormhole things, you can use this one item to teleport here once every x hours, etc. etc. Kind of a one-stop guide to minimizing down time in a, say, four hour archaeology shift, and maximizing the most digging you can do.

[mupto]

your time+wow=blizz gets rich?

is this the math you speak of?

[YersiniaPestis]

I don't think Archy is meant to maximise your fragment per hour count. It just gives you information.

Also, it depends on the fact if you are aiming for a particular artifact, for example, bug mount. In that case Uldum digsites may have higher priority over any distance.

Although, after recent patch, if you only have one rare elven artifact left to make, it could make sense to just go for the elven dig sites in order to quicker make that rare artifact and then get mostly uldum dig sites.

[Deleted]

oh really.........