I saw this interesting video lecture by Nassim Taleb. He is involved in the financial industry, currently an advisor for the IMF, and the author of, among other works, The Black Swan, a book about the concept of fragility. Here is some of what I took away from the ideas presented in this lecture and maybe how it might relate to competitive gaming. An important disclaimer: I haven't read the book and I'm not an expert on the subject matter (I feel like I can get away with this in a blog ) so if you want to know more you should read the book, find a summary information somewhere or watch the video and so on, and above all, don't take my word for anything.
A black swan event is defined by Taleb as an outlier, a rare and unpredictable event. His notion is that much of the history of the world is shaped by events that could not have been predicted beforehand. The obvious takeaway is to build a society that can benefit from these events, that is robust in terms of dealing with them, rather than to try to create a model that can predict the unpredictable and to find safety in that manner.
The unpredictability is due to certain factors, Taleb names: "the impossibility of possessing all relevant information, that small unknown variations in the data can have a huge impact, and flawed theories/models that are based on empirical data and that fail to consider events that have not taken place but could have taken place." Furthermore, unpredictable events will have larger impact by virtue of lack of preparation for them.
A key idea is 'fragility'. Think of porcelain: if you let it fall then it breaks and there's nothing you can do about it anymore. Taleb has a general rule which is that "a negative event of size X is more harmful than a negative event of size X/k done k times". Meaning that harm is non-linear, you won't survive falling off a cliff, but you will survive jumping up and down a thousand times, even if it's the same distance in the end. And everything that is alive will have this non-linear harm distribution, simply because otherwise it could not survive the environment it lives in. And this means that a failure to predict something of severe negative impact results in death.
Systems that are fragile are vulnerable to time, since time will eventually kill all fragile systems. (it just takes one negative event of sufficient power) This makes them fearful, prone to centralization and attempts to model and control the entire world, enterprises that are unfortunately doomed to failure and that are actually reinforcing weaknesses as more centralization leads to even more fragility. Taleb advises to not try to create a model of the environment, which is too difficult, but to try to understand how individual components can relate to the environment, which is a smaller and more manageable task. In the latter case, more small scale interactions with a larger framework offer you an opportunity to 'experiment' with more extreme values as well.
Taleb distrusts top-down knowledge, amusingly enough he calls for the abandonment of the Nobel Prize for Economics, stating that top-down economic theories are potentially devastating and that most actual knowledge gained is due to 'tinkering' by people that work in the business, rather than by academics who can at best formalize some of the ideas.
From what I understand, Taleb prescribes the following dictum: create an anti-fragile system by taking more risks. That seems counter-intuitive, but it's based on the idea of "what kills me makes others stronger" (Taleb's phrase). A bridge that is destroyed due to some event makes other bridges safer because of more knowledge gained for improved safety protocols. And logically, if the weaker elements die the stronger faction survives. If individual components are allowed to fail 'sequentially', you have an overall stronger system. And if not for risk taking (Taleb writes mostly about the financial world, so his example here is encouraging entrepreneurship), the differences between the components of a system are not meaningful and there is no actual improvement in the system.
Here is the video again, so you can watch for yourself:
--
I think everyone will have different opinions on how this relates to the Starcraft 2 scene. It's presumably easy to find evidence in here for your favorite pet theory on how to improve the scene, but in any case, what I would take away from it is to encourage grassroots development and to not put all our faith in Blizzard or KeSPA. A scene sustained by local interest and initiative will probably survive for longer than one under the iron fist of institutions like those two. Blizzard can become bankrupt or simply decide that SC2 is no longer profitable, the Korean government can decide on new policy and so on, so it's really quite dangerous.
On the other hand, the original Korean e-sports scene is a perfect example of anti-fragility: many games came and went, but Brood War captured all the ingredients necessary to service a burgeoning e-sports scene. Nobody was harmed by games that failed, but the Korean society was in great position to benefit from a game that was successful.
--
I think the book was quite famous, so probably someone knows more about it than me, so please correct me on things I didn't put correctly, thanks.
On September 30 2013 02:17 farvacola wrote: Nassim Taleb is a cool guy, and you did a pretty god job of relaying some of Anti Fragile
I thought it was an interesting video and doing this write-up helps me understand the ideas more so I encouraged myself to write this. Maybe it's a bit forced to try and relate it to Starcraft 2 though, I had counted on being able to come up with some relevant ideas inspired by Taleb's video but that section ended up smaller because I was only going to repeat what others have said already anyway. :/
This is interesting. Thanks. I'll watch the video when I get home.
Taleb appears to base his research on work by Hayek and other systems thinking. This is especially if I take your highlighting of the dangers of centralization and top-down thinking. This is because of the central problem of knowledge within complex systems. I blogged about it myself not too long ago: http://www.teamliquid.net/blogs/viewblog.php?topic_id=429165
I applied, or tried to apply it to SC2 design and balance concerns. As you know, I've long advocated a laissez faire approach to these issues in SC2. I prefer if Blizzard leaves it to players and map-makers. To my mind, constantly appealing to Blizzard to "fix" the game runs the danger of ruining it entirely. I also believe the game is intrinsically healthier and better than any one individual thinks it is. Posting about Hayek's ideas and how they relate was a way of beginning to organise my thoughts.
I believe the same may be applicable to organization of SC2 at grassroots level - as you seem to suggest. Knowledge of players and fans in the regions and at the margins is more available to providers in those locales than it is to Blizzard. The danger is not so much that Blizzard will become bankrupt or the Korean government change policy than in Blizzard not having sufficient knowledge in sufficient time to make the best decisions possible for the betterment of the game. However, this does mean that Blizzard must be willing to cede control to providers in the regions and the margins.
Hayek is mentioned in the video also, by the way. (during Q&A)
And well, one of the points of (anti-)fragility is that the world isn't a game, so you don't know the rules and you have less power to predict outcomes that way. However, maybe Taleb is referring to games like Poker and Chess, which are more comprehensible for academics involved with game theory and such. Starcraft 2 is quite different and is more like a 'complex, multi-variate system' (to quote Hayek) and furthermore is strongly limited by the player's lack of imagination and mechanical ability. The game essentially reveals itself to us through the actions of the players so on some level properties the game has aren't innate but only exist (for relevant purposes) once they see play. In any case, I don't know if any of the specifics of the main argument are easily applicable for game design, but some of the associated thinking certainly is.
You might know my pet peeve of disliking the concept of "fundamental problems with the design of the game" that you see floating on reddit&TL in game design discussions often. It's usually about warpgate, which breaks defender's advantage and so on. My belief is that there is no blueprint you can make for a game, all these rules that the community comes up with like defender's advantage are only guidelines and any game worth its salt should not try to mindlessly affirm this in every element of the game but should try to sometimes challenge it to increase variety and depth. Every asymmetrical piece of game design will have some tension and some hidden explosive potential which if not contained can be exploited to break the game. I personally think that warpgate is broken, but it's because every unit is potentially broken and it depends on the context of the game whether it's deemed truly problematic. The context of the game consists of all other units (i.e. it's somewhat arbitrary) and also the knowledge of the players. In the end it's a statistical argument that's much like "is this enjoyable enough of the times?", so it comes down to: "does this work in the game?". You don't need to be a genius to notice that warpgate often leads to problems, but it's dangerous to act like it has to do with anything fundamental.
I think this means that to be a good game designer you basically don't need any knowledge, you just need cleverness and perceptiveness, the ability to come up with creative solutions and so on. The team liquid forum is essentially unnecessary for anything other than signaling what aspects of the game are disliked by players, which is useful information for Blizzard, but something they can also find from other sources. There is a delusion among many people here that they understand the game and would know better. However, you can't completely predict how changes are going to affect the game, (or how the meta is going to change), it's too complicated. In the end only Blizzard has the proper resources to do any in-house testing and they're obviously the only ones that matter in terms of what changes are made, so it would be nice to have more humility in the game design discussions instead of all this campaigning which is so typical of reddit and seeps over to TL.
But yeah, probably the take away for game design is to not obsess with constant tiny changes, but rather to design in such a way that you don't need to always do that.
After a professor got me Taleb's Anti-fragile after I finished working with her, I always wanted to write up something like this. If you haven't read the book, I very much encourage you to, it's very well written.
There' are other pretty neat things that it helps you think about (PoE design vs. Diablo 3 design, LoL vs. DotA 2, Origin vs. Steam, or even Mac vs. PC).
EPT
) so if you want to know more you should read the book, find a summary information somewhere or watch the video and so on, and above all, don't take my word for anything.
