17,094 reputation
45282
bio website thenerdiestshirts.com
location
age
visits member for 5 years
seen 6 hours ago

13h
comment Asymmetric random walk on the line with barriers
One simple observation is that if the distribution is supported on $\mathbb{N} \cup \lbrace -1 \rbrace$ then it can't be an example, since there would never be any overshoot in the negative direction yet asymmetric distributions would have some positive overshooting. Similarly, the distribution has to include a positive probability of taking a positive step of size at least $2$.
19h
comment A possibly surprising appearance of Lucas numbers
Instead of using $x$ or $r$, why not just use $\sqrt{2}$ everywhere?
1d
comment Genus of Covering Space of 3-Manifold
See mathoverflow.net/questions/86800/….
Jan
24
comment optimal strategies for 2-player zero-sum games of perfect information
Are you asking for something other than Zermelo's theorem?
Jan
23
answered Is mean width a Dehn invariant?
Jan
20
awarded  Enlightened
Jan
20
awarded  Nice Answer
Jan
17
comment Bayes with non-parametric data
There is no problem with using Bayesian techniques with log-normal data. That's still parametric in a different family. It sounds like you are asking whether something might go wrong if you use the wrong model. Of course the answer to that is yes, but sometimes you don't expect any noticeable differences.
Jan
16
comment Does bounding moments make distributions close in total variation distance?
@Bullmoose: One possibility would be to impose conditions such as that the density is unimodal. Another is to look at a different distance than the total variation distance.
Jan
16
answered Does bounding moments make distributions close in total variation distance?
Jan
16
awarded  Necromancer
Jan
13
comment Probability that top k elements will be got?
Variations of this have been asked before with Gaussian noise, no delta mass at the original values, and original values not in an arithmetic progression. I don't know that there is any simple answer, but how attached are you to this particular model? If there is no complete formula for everything, then you might need to focus on some part of the range of values. Which region is of particular interest?
Jan
11
answered How to roll a $p$
Jan
9
comment covering high dimensional hypercube by balls
@usul: For $r=1$ it only makes a difference in low dimensions.
Jan
9
comment covering high dimensional hypercube by balls
@usul: That lower bound (for $r=1$ before the OP changed the radius to be arbitrary) is not correct for the version where the centers can be arbitrary. If $d=3$ and $r=1$, all corners of the cube are contained in a unit sphere centered at the center of the cube.
Jan
8
comment covering high dimensional hypercube by balls
I would be surprised if the case of arbitrary centers has been studied much.
Jan
8
comment covering high dimensional hypercube by balls
Are you letting the centers of the balls be points other than vertices? The corners of a square are contained in a unit ball around the center but not any vertex.
Jan
8
comment How large must $A$ be if $\{1, \ldots, N\} \subseteq A-A$?
Allowing negative numbers is unimportant because you can translate $A$.
Jan
4
comment PDF of th product of normal and cauchy distributions
Are these centered?
Jan
3
awarded  Yearling