Page 1 of 1

CosmoMC: SDSS Fitting Code

Posted: May 24 2006
by Jun-Qing Xia
Dear everyone,

I have a question regarding the fitting code to SDSS power spectrum of CosmoMC.

In the May 2006 version seems the following part has been revised again:

Code: Select all

  w=1/(mset%mpk_sdev(:)*mset%mpk_sdev(:)) 
  normV = sum(mpk_WPth*mpk_WPth*w) 
  tmp=sum(mpk_WPth*mset%mpk_P*w)/normV 
  chisq = sum(mset%mpk_P*(mset%mpk_P - mpk_WPth*tmp)*w + log(normV) 
where the last part 'log(normV)' is the same as the pre-WMAP3 code, but differs from the March 2006 code where this term has been dropped.

I am puzzuled about this.

Thank you for any suggestion.

Jun-Qing Xia

Re: CosmoMC: SDSS Fitting Code

Posted: May 24 2006
by Antony Lewis
I accidentally commented this out in the March 06 version - it does say in the version history that this was a correction

http://cosmologist.info/cosmomc/readme.html#Version

See appendix F of astro-ph/0205436 for discussion. (the log term generally does not make much difference, and depends on the prior)