Loss-of-sales is a bit harder to quantify, but it's touched upon by a comparison basis by this article: http://www.gamasutra.com/view/feature/3030/keeping_the_pirat...
I don't extend my claim to music or any other form of infringement.
[0] http://gamedev.stackexchange.com/questions/54749/has-piracy-...
Had the pirated copies been sold - the same server issues would have happened. Though they would have had more resources to handle the situation and it would turn out better for everyone (especially the iOS devs who had to go offline for 4 months to deal with the situation).
However, "offline" games (ie. solo campaigns) or distributed servers (ie. minecraft) that do not connect to a centralized server won't take away company resources. The idea behind piracy is that nothing is being stolen or taken, only duplicated. Taking away server resources has a negative impact and although it is also considered "piracy" I would consider it closer to "theft" (the same as stealing a physical server might be!)
Without arguing over semantics of the meanings behind the word 'piracy' - will you agree with me that the shared issue of all of your examples was unpaid server space being 'stolen' by pirates? We may disagree on whether lost sales matter or don't matter in this scenario, but I hope we can both agree that the stolen server space was definitely an issue.
The Gamasutra article where money was spent on making piracy harder doesn't give any metrics for a meaningful conclusion to be drawn.
Did Spyro 2 being cracked in a week actually harm sales at all? Difficult to quantify. How about the profit from the next Spyro? Was it more or less profitable? If it was less profitable - was that a result of spending money on piracy protection? If it was more profitable - was that a result of adding piracy protection or because the Spyro series had surged in popularity after the success of Spyro 2?
The answer is difficult to quantify, especially without relevant metrics given. What was the cost of adding the piracy protection? Can you quantify "popularity growth" between the two games? Can they estimate how many people would have bought the game if they could not pirate it?
Some game devs spend a lot of time and money trying to prevent people from modding their games. Some game devs make it easier to mod their games or even encourage it.
If a mod gets popular and gets a community growing around the game - is it doing good or is it doing harm? Assuming that a large amount of the community growth is from people pirating the game to play the modded version (and assuming none of the above server issues)
There are 2 sides with difficult-to-measure metrics.
1) Is the mod and the pirates playing it bringing enough attention to the game and free publicity/advertisement as to onboard new paying customers?
2) Is the time and money spent on preventing such a mod community from growing going to cost more than potential "lost sales" of pirated copies of the game? In addition, what is a "lost sale" and how are they measured?
To expand upon the issue of "lost sales" a bit:
Plenty of pirates argue that if they could not pirate a game - they would not purchase it either. I fall into this crowd. Outside of the stupid amount of money I've spent on League of Legends, I've only spent money on games I originally pirated. I purchased the game, purchased DLC, or purchased merchandise related to the game. If I could not pirate the game - I would not bother playing it, telling my friends to buy it, or buying merchandise from it (since I wouldn't know I liked it!)
By allowing people like me to pirate you actually win an otherwise lost sale. This is what makes these metrics so hard to quantify. Too many game studios consider all piracy to be a lost sale when people like me are the exact opposite. Without piracy we would have been a lost sale.
So it comes down to "which game studio/game dev do you want to take their word for?" Many of them attribute their success to piracy of their games. Some of them say piracy killed their game.
Without quantifiable metrics... how do we know the "killed games" didn't die simply because they were bad games that, piracy aside, nobody wanted to purchase?
From a business perspective, that's the far more likely scenario than the good games that the game devs claimed piracy helped selling well regardless of piracy because ???
TL;DR
My argument is that piracy doesn't kill bad games due to lost sales. They don't get any sales because their game is bad and few people are willing to pay for it. On the contrary, popular pirated games still do amazingly well because they are good games that people are willing to pay for even with the availability of pirating. To myself, makes more sense from an economic perspective for gaming media.
Music pirates in particular tend to buy more music than non-pirates. I would wager this holds true in any media-field of pirating (but perhaps not software pirating). [0] There is a reason for this!
Pirates are enthusiasts. They love the content they consume and are likely to pay/buy things they really enjoy/enjoyed. Due to not having an unlimited income they pirate some things. For music pirates - there is a good chance that the music they are purchasing is from the same music label as the music they are pirating.
Pirates are also likely to tell their friends about a show they enjoyed. Pirates are also likely to buy merchandise from shows they enjoy.
For example - an extremely large portion of the American market for Japanese Anime-related goods is because of pirated fan-subbed anime. Shows that would have never been released or heard of in America and thus led to 0% revenue from America are pirated and their merchandise purchased. The number of figurines from おれいも for example. Or the number of だきまくら purchased from American-fans.
This simply can't be denied for the anime market in particular - and many fansubbers eventually get hired by the studios. Some studios try to shut down the fansubbers but most allow it to happen because fansub groups largely do not sub anime slated for an English release. It's an unspoken code of sorts of "Support the official release" - and since many anime fans don't necessarily speak Japanese, if someone isn't translating it for them they can't/won't watch it.
Pirating is beneficial for the anime scene. It's likely more beneficial than harmful for music for similar reasons. I'll concede that for Movies/TV Shows it gets a little iffy (do enough people buy merchandise who wouldn't if they hadn't pirated the show to see it?)
Then we can also look at video games. Amnesia: The Dark Descent or Minecraft. Games that were heavily pirated and the creators even went so far as to say "If you can't afford to play our game. Pirate it." Why? Free marketing and free publicity means more sales. It made economic sense to let your game go viral that way. What might be $20 in a lost sale to a pirate might win you $80 in 4 new sales from his friends (even if another 6 of his friends just pirated the game)
If you don't let him pirate the game you just lost $80 in sales and free marketing/publiclity from 7 people which likely lost you even more in sales.
Which brings me to the important part: Pirating doesn't mean a lost sale. It means free marketing and free publicity from someone who wouldn't have been a customer. If you can convert a "non-sale" to a "promotional consumer", why wouldn't you?
The economic spin usually given to pirates is it is "costing sales" which is, frankly, bullshit. They make more money off suing mass pirates than any "lost sales", so it's beneficial for them to keep pirating illegal.
[0] http://www.huffingtonpost.com/2013/01/22/music-pirates-study...