Von Clausewitz a Bayesian? - May 15, 2008
"Our knowledge of circumstances has increased, but our uncertainty, instead of having diminished, has only increased. The reason of this is, we do not gain all our experience at once, but by degrees; thus our determinations continue to be assailed incessantly by fresh experience; and the mind, if we may use the expression, must always be 'under arms.'" On War, Carl Von Clausewitz
I like the Bayesian notion that we must constantly be examining our hypotheses against new information. And, if we take it a bit outside the math context (where I am more comfortable), it is our job to find that data. Because as Steven Landsburg put it, under Bayes' Law "everything that can be relevant is relevant."
A theory about the world or who you want to be - combining Von Clausewitz and Bayes - is a bit like a battle strategy. It doesn't mean anything until it comes under siege. The enemy nor the data cares much for your plans. Success requires a certain prescience, sure, but more crucially, it needs a general that closely monitors the feedback from all sources of information and consistently learns from them.
And I don't mean this is in the generic "can you be flexible?" sense that everyone throws around. We're talking objectively, can you monitor and track your actions empirically? How quickly can you rewrite your operating procedures? Which of your assumptions are firm and which can be shifted? Do you fear new information or do you welcome it? How active is your pursuit of challenging feedback? Have you identified the false positives so you can avoid them?
From Clausewitz, I am thinking that difference between someone who can launch and learn and someone who can't isn't so much a difference in skill level as it is of planes.
Posted by ryanholiday at 4:40 PM
Print Friendly · Digg it · del.icio.us · StumbleUpon · Netscape
Trackback Pings
TrackBack URL for this entry:
http://www.festeringass.com/mt/mt-tb.fcgi/2376
Comment Policy:
Anonymous comments are allowed. All anonymous comments and comments from those not registered with TypeKey are moderated. They WILL NOT appear until they are read and approved by a moderator.
It is strongly encouraged that you sign up and login with a TypeKey account. Once you do that, your comments will be immediately posted.
Comments
Way to go Ryan, this post on the relative strategy and false positive strategy is exactly what I've been thinking for the past month. you hit this one out of the ball park!
Posted by: Eric at May 16, 2008 09:44 AM
Bayesian probability is pretty sweet, but not without its problems. I'd suggest a pretty clear article on its benefits and shortcomings vs frequentism here: http://scienceblogs.com/goodmath/2008/04/schools_of_thought_in_probabil.php
Posted by: francisco at May 16, 2008 01:43 PM
You think because you go to Berkeley you know more about probability than me? I USED WIKIPEDIA!!!!!!!!!!
Posted by: Ryan Holiday at May 16, 2008 03:30 PM
I go to Cal and use Wikipedia and this offends me a lot.
Posted by: Avinash at May 16, 2008 04:44 PM
Seriously, this blog has gone downhill. I must have gone Hollywood or something. Everyone knows Wikipedia caters to the lowest common denominator.
Posted by: Ryan Holiday at May 16, 2008 05:29 PM
Ryan, stop bitching about wikipedia and focus on your intellectual endeavors such as Fail Dogs.
Posted by: Gris at May 16, 2008 06:01 PM
Woof.
Posted by: Andrew McMillen at May 18, 2008 12:26 AM
Post a comment
RECENT ENTRIES
Things I don't like about myself since I moved to LA
July 3, 2008
Bad Moments in Online PR or (Most People Suck and That's Great News)
July 2, 2008
