Sunday, July 8, 2012

Diablo 3 Crafting: Inefficient and Stupid

When Blizzard calls something a "gold sink", that's code for "don't fucking do this, ever."

Diablo 3 offers players the opportunity to craft items.  One may salvage useless magical and rare items into crafting materials that are then recombined to form new items.

I initially declared crafting to be inefficient just by looking at the numbers.  Some other players, however, maintained that crafting is a sensible investment that can pay off over the long run.  They would post pictures of the spectacular items they had crafted and the astronomical amounts of gold they made doing so.  Yet when I asked how much of that was profit, and how much of that covered their expenses, they were speechless and incapable of discerning the meaning of the question asked.

So, I decided to toss a few million gold down the sink in order to compile the data required to:

1)  Show people how to fucking play Diablo 3.
2)  Declare crafting to be inefficient and stupid.

Here's the spreadsheet:

On the left is a list of my total expenses.  I recorded the gold I spent leveling Blacksmithing, the gold I spent on a pattern, and the gold I spent crafting items.

The middle column under "glove construction" is used to discern the cost of every pair of gloves.  I input the daily Auction House list price of essences, tears, and tomes, and the spreadsheet calculates the cost of one pair of gloves.  Below that (not pictured) is a cell where I input the number of gloves I made, and below that is a cell that displays the amount of gold I spent.

Moving to the right we see the Gloves Sold column.  I record the name of every item, and if you mouse over the cell an image of the item pops up (so I know exactly what the item is).  Then we have the price each glove sold for.

Finally, we see the total profits and losses.

Let's look at the numbers.

Total Expenses:  4,025,503
These expenses include:

- Leveling the Blacksmith:  You have to factor in the gold you spend to be able to craft in the fist place.
- Pattern price:  Since I had previously sold all the crafting patterns I obtained, I had to buy a +5 affix glove pattern.
- Crafting 38 gloves.

Overall, I spent 4,025,503 on crafting.  If we remove the cost of the BS leveling and the pattern, I spent 3,023,248 just on crafting gloves.  Those are the expenses, let's now look at the sales:

Sales:  1,912,500

My total "profit" from selling 5 pairs of gloves is 1,912,500.  Only one pair sold for over 1M, so the rest are 100K - 300K sales.

Now, if you look at just those sales, crafting seems to be profitable.  After all, I made 1,912,500 gold from selling gloves.  Not bad, right?  Woo crafting!

As with any other business, though, you have to compare your expenses with your sales.  So, let's do that.

Total expenses:  4,025,503
Total Sales:  1,912,500
Total Profit:  -2,113,003

My total profit is a loss of 2.1 million gold.

Just to rub some salt in the wound, let's look at some more numbers.  Suppose that I had never leveled my BS or bought that pattern.  Suppose I had never crafted those gloves.  Finally, suppose that I had just sold the crafting mats on the Auction House, rather than use them to make shitty gloves.  How much would I have made in that world?

To figure that out, we take the number of each crafting material and multiply it by its sell value.

Essences: 760 * 1600 = 1,216,000
Tears:  152 * 170 = 25,840
Tomes:  228 * 1100 = 250,800

If I had just sold the crafting mats, I would have made 1,492,640 gold.

Granted, that 1,492,640 is less than the 1,912,500 I made by selling gloves.  However, once we factor in the crafting expenses, we discern that selling the crafting mats is a far safer investment that always pays off.

So, TL;DR:
Total Expenses:  4,025,503
Glove Crafting Expenses:  3,023,248
Total Sales:  1,912,500
Total Profit:  -2,113,003
Crafting Mat Sell Value:  1,492,640

I've heard tell that persons who stumble upon +6 affix crafting patterns and craft hundreds if not thousands of items can eventually turn a profit.  You craft 100 +6 hats, and maybe generate one that sells for 20M+.  If you do that enough, you can start to make some money.

However, most people do not have the venture capital required to make that initial crafting investment.  Moreover, the market is constantly changing with a wide disparity between the price of ideal items and the price of shitty items.  The vast majority of your crafted items will list for 5,000 gold, and no one will buy them.

If you got into the crafting market during the first weeks of the game, then you could have made some money.  If you're sitting on a pile of hundreds of millions of gold, and want to take the risk, you might be able to turn a profit after you grind out a thousand or so items.

But for most players, if you actually look at the math, crafting is completely inefficient and stupid.


Mike Lewis said...

Are patterns one time costs or recurring?

_J_ said...

Patterns are a one time cost. Obtain the pattern, teach it to your artisan, done.

Mike Lewis said...

Okay - so your your math is missing a step.

if d3 were a true economic system then training and patterns are have to be placed in a second expenses column: They are capital expenses. They are a cost you have to eat in order to make a product. Ford has to spend millions of dollars to build a new factory. Those are one time costs. the parts to make the car are recurring costs. you will eventual recoup the costs of the capital expense.

Over time, the balance will change. If you have a good pattern that produces an Ax with attribute that are in a high demand - you will make back to cost of training/pattern.

But as you and everyone else has shown - d3 does not have a true economic system. there is an unlimited supply of gold and crafting supplies.

_J_ said...

"if d3 were a true economic system then training and patterns are have to be placed in a second expenses column: They are capital expenses. They are a cost you have to eat in order to make a product."

A few people have said this, and I think they're incorrect.

I placed the pattern in the expense column. It is an expense that I hope to offset by profits.

However, I have not yet made enough profit to pay off that initial expense.

So, in my reckoning, it goes in the expense column with everything else.

Why is the pattern expense supposed to be treated differently than, say, the crafting mat expense? They're both

- gold I spend to produce gloves.
- expenses that my profits are supposed to overcome.

It seems like people don't want me to factor in the capital expenses...and that seems to be obviously silly.

If I hadn't bought the pattern, or trained Blacksmithing, I'd still have that gold. So, why isn't it an expense?

_J_ said...

Capital expenditure:

"For tax purposes, CAPEX is a cost which cannot be deducted in the year in which it is paid or incurred and must be capitalized. The general rule is that if the acquired property's useful life is longer than the taxable year, then the cost must be capitalized. The capital expenditure costs are then amortized or depreciated over the life of the asset in question."

Yeah, I don't think that applies to Diablo.

Mike Lewis said...

They are patterns and training are an expense - just a different sort of expense. You pay for those once. The crafting supplies are a recurring cost.

How are the different? When crafting, you want to off set to cost of supplies with each ax you make where you are paying down the cost of the training/patterns over time.

kylebrown said...

While I agree with your premise, this is a painfully small sample size.

Were there really only 5 pairs of gloves made, or were there trash gloves made also? If so, were those trash gloves vendored or parted out for crafting again? Did you include those assets in your final profit margins?

_J_ said...

"Were there really only 5 pairs of gloves made, or were there trash gloves made also? If so, were those trash gloves vendored or parted out for crafting again? Did you include those assets in your final profit margins?"

Gloves made: 38
Gloves sold: 5

So, 13% of my gloves sold on the AH.

I salvaged the shit gloves in 1 essence, 1 tear. I did not include those in the price, but it would not be difficult to do so. 33 essences and tears still do not offset my losses.

I admit that 38 gloves is a very small sample size. But, at the moment, I'm not willing to invest the gold required to increase the sample size.

_J_ said...

"They are patterns and training are an expense - just a different sort of expense. You pay for those once. The crafting supplies are a recurring cost.

How are the different? When crafting, you want to off set to cost of supplies with each ax you make where you are paying down the cost of the training/patterns over time.

I get that the pattern and training are a one time expense. But they are still an expense. At the end of the post, I give detailed expenses:

Total Expenses: 4,025,503
Glove Crafting Expenses: 3,023,248
Total Sales: 1,912,500
Total Profit: -2,113,003
Crafting Mat Sell Value: 1,492,640

The 4,025,503 includes the pattern / training. The 3,023,248 is just the crafting materials. My sales were 1,912,500. That does not cover either the crafting costs, or the crafting + training + pattern costs.

I understand that capital expenses are treated differently for tax purposes.

I don't know why people want me to treat them differently in Diablo.

Because a capital expense is still an expense, just like the expense of crafting each glove.

_J_ said...

Two other things to consider / talk about.

1) What constitutes an adequate sample size?

How many pairs of gloves would I need to craft in order to discern the profitability of glove crafting? Given how random the stats are, is it even possible to generate a sufficient number of crafted items to discern profitability?

2) The definition of "good" changes.

As the market shifts, players leave, and characters gear up the threshold of "good" changes. A month ago more of my gloves would have sold. A +50 str/vit pair of gloves with some attack speed was awesome around the time of release. Now it's shit.

So, when we talk about sample sizes and adequate data, we also have to factor in the changing criteria for what will sell. As time goes on, less items sell, due to:

1) Persons wanting higher numbers.
2) The shit market flooding.

This isn't like poker, where there is a defined heirarchy of hands that beat other hands. This is a changing market. That pair of gloves I sold for 1M last week will be maybe 500K two weeks from now.

So that's another consideration to factor in.

Mike Lewis said...

Crafting is bullshit. Not saying that it isnt. It can only be effective at scale - so making 100s of gloves.

I framed this whole thing in terms of a "real economy" which diablo is not. Its a set of mini games that have no pay off.

_J_ said...

"Its a set of mini games that have no pay off."

I've made $295.80 so far.

That's a payoff.

Or, it's more of a payoff than I ever got from WoW...or D2...or any other video game, come to think of it.

Caleb said...

But wouldn't that $295.80 be nothing compared to, for instance, a $50,000 10 month contract teaching at a selective liberal arts college, which you would have a real chance of acquiring once you pass your prelims and successfully defend a dissertation?

Puri said...

I'd recommend leaving opportunity costs not directly related to playing Diablo out of this, because I don't know of any way that it would not instantly end the conversation with "wait, fuck this game entirely".

MA17 said...

That was me, just typing from a different account and getting unexpectedly different results.

I agree that handling capital expenses isn't necessary if you're trying to answer the question "have I made a profit from crafting?"

However, to build on Kyle's point about the small sample size, if you want to answer the question "Is it possible to make a profit from crafting?", then there's more that needs to be done. More items need to be crafted, and capital needs to be separated from operational expenses.

The basic idea is that you need to determine how much gold it costs to craft one item (not including pattern and BS leveling). You have fixed overhead in the form of the BS fee (42,416G), which will (I assume) only change via patch from Blizzard. Then you have variable overhead in the form of mat prices. It cost you something like 39,280G in mat costs in your example, bringing your total gold cost for one crafted item to 81,696G. Fixed/Variable difference is relevant because judging from your expenses column, it looks like your mat cost has varied by as much as 5,000K. When we start looking for ways to make crafting profitable, controlling your variable overhead will be pretty important, so it must be evaluated separately to be useful.

At this point we know what crafting costs and that we have to charge more than that cost in order to make any profit. What we don't know is how much we need.

For something like this, we should look at your defective rate next. This is the number of items you discard versus the number you sell. As you show, 5 items sold out of 38, meaning you sell only 13.16%, or 1 out of every 7.6 crafted. That means you have to make, on average, enough gold to offset the cost not only of the item you're selling, but of the 6.6 you're not. That works out to roughly 621K gold(81,696G * 7.6 = 620,890G) in order to approximately break even on a successful item sale, on average. To make profit, you have to charge even more. You're running between 127,500G and 1,020,000G here across 5 sales, averaging 382,500G per sale. Compared to the 620,890G you need to average, you're not profitable and you never will be, based on the available data.

Which is why a larger sample size will be useful here. It will give you a more accurate picture of your defective rate, and allow you to determine if your sales even have the POTENTIAL to be profitable, which ultimately means you have recouped the capital expense of the pattern and the BS leveling.

As it stands, you're absolutely right, what you've done has not been profitable, nor does it seem that it will ever be. However, if you can find ways to get better rates on mat, improve your defective rate, salvage your scrap (sell for cheap, recycle into mat? not sure what would be best there), and then calcuate based on a larger sample size to see if crafting can be profitable.

MA17 said...

Also, the sampling may not be necessary if it's possible to know the probabilities determining which valuable stats will be on your crafted item, and can get reasonably accurate data on what items with those stats are going for at the time you're crafting.

Might not hurt to be crafting a variety of items in case pants happen to be in high demand for some reason.

Caleb said...

It seems to me that MA17's analysis is spot on, although, I find myself drawn to the concept his alter ego threw out: "Wait, fuck this game entirely."

_J_ said...

"Wait, fuck this game entirely."


Right now I'm at the point where...ok...

You know how in sitcoms when a parent finds their kid smoking they force them to smoke a carton, and the resulting pain and sickness makes the child never smoke again?

I'm the kid. Diablo 3 is smoking. And I'm closing in on the "Daddy, can I stop now?" point.

Which is really good timing, since I have to go to a conference next week and won't be able to play.

I've said to a couple friends that Diablo 3 may be what makes me stop being a gamer.

That seems more likely the longer I play this fucking train wreck of stupid.