MCS Is a great and necessary tool that’s in virtually every options trader’s toolbox.

One can run 500,000 samples or runs in seconds or minutes. I’d be glad to discuss this to you if you’d like. My other email address is biobot@gmail.com. I check that one every few minutes. Ethermadness is more for private communications to select colleagues.

Good luck on you study. Wiki the subject if you haven’t already. Wiki is great for current thinking and uses on MCS. It’s way more current tha any dead tree writings by its nature.

Btw I’m a math and physics freak turned quant in training. I’m going to check out that book if it’s cheap enough.

If the model is not realistic, getting more precise estimates from it by running it longer (or with more random starts) may not really reduce risk. (I guess you can try to do additional computations, however, to try to assess the accuracy of your model and maybe account for it.) I guess you usually don’t go wrong, however, by adding more data, and the computation required naturally also goes up with the amount of data.

> Usage, credibility and diffusion would be interesting, but would require very different strategies to assess.

Are those technical terms? Do you have any operationalization in mind, or hunches?

> 1 and 2 differ only in network degree themselves

I agree, except for medium, as you point out. But maybe to simplify we could start with social media data and use degree as a proxy for mainstreamity. See https://twitter.com/cnn for example.

I hear Vern Paxson is sitting on a huge pile of Twitter data.

> 3 and 4 also only differ in terms of how the steps are organized (i.e. who gets the last laugh?)

Not sure I see what you’re getting at. Seems like for 4 there’s a bit more direct solicitation of user input (enticing people to click the Like button is different from PageRank that works on linked content?)

> we are left with identity and filtering strategy (and its transparency?)

Can you elaborate on what you mean by this?

3 and 4 are both combinations of 1 and 2 (see PageRank, social search, etc.). 1 and 2 differ only in network degree themselves. 3 and 4 also only differ in terms of how the steps are organized (i.e. who gets the last laugh?).

At least we are simplifying the independent variables – we are left with identity and filtering strategy (and its transparency?)

What kind of metrics could be used to compare those methods, though? They seem to be quite different processes:

1) organic (strategic?) propagation through a network

2) broadcast through a narrow channel (now with substitutes, but presumabely still important because of some combination of market inertia and its ability to reproduce the conditions of its own success through funding high-quality content)

3) direct search (which begs the question, why are people searching for what they are searching for in the first place?)

Here’s an argument: (4) is actually much like a combination of 1-3, in that they are often facilitated by algorithms, and harness social network effects in order to curate content for a narrow distribution channel to a large audience.

It’s interesting to think of the consequences of that. One of the problems with PageRank is that it is possible to outsmart and manipulate it. Would spammy academics be able to use a similar strategy?

http://www.google.com/url?sa=t&rct=j&q=how%20to%20attack%20pagerank&source=web&cd=1&ved=0CBwQFjAA&url=http%3A%2F%2Fwww.eecs.harvard.edu%2Fcs286r%2Farchived%2Ffall08%2Ffiles%2Fne06-sybil-1.pdf&ei=SbvbTreJEqihiQLf46zoCQ&usg=AFQjCNHSktmXXLc14mbmiD6aglnarz8pBw&cad=rja