Study of the dijet invariant mass distribution inThat's despite the fact that the CDF claim grew to 4.8 standard deviations - formally approximately 2 parts per million risk of a false positive - before it dropped to about 4.2 standard deviations (27 part per million risk of a false positive). Many people were almost certain that the signal had to be real.

pp → W(→lν) + jj final states at √(s) = 1.96 TeV

D0 has analyzed 4.3 inverse femtobarns of their data. They formulated their opposition to CDF in this way: the probability that their (D0) signal is compatible with the CDF's claim of a dijet resonance near 145 GeV that has the cross section of 4 picobarns is just 0.000008, i.e. 8 parts per million.

Because D0 has obtained an agreement with the Standard Model, using the same data as the original CDF paper, and I believe that it's more nontrivial to get an exact agreement than to introduce a random error and inappropriate bump, I am almost completely convinced that D0 is right, CDF is wrong, and there is no bump. As a jury member, I say: Abazov beats Aaltonen! ;-) This expectation - that conservative physicists are almost always right and the "progressive" ones are almost always wrong - holds for the same reason why most big mutations of an organism are detrimental.

If you will hear someone saying that near-5-sigma "proofs" are incontrovertible, tell him or her about this story!

*The CDF Bactrian (2-hump) camel is actually a D0 Arabian (one-hump) camel i.e. a dromedary. The 150 GeV hump was a Fata Morgana. Picture of D0 via Dorigo: click to get to his blog.*

It's hard to figure out what CDF is doing wrong. But the evaluation of the jets and leptons is very complicated. If it were possible, the D0 people should try to apply their algorithms on the CDF data even though it's not quite possible because even the raw data differ because of the different detectors.

Needless to say, the likely CDF debacle is making their top quark-antiquark asymmetry doubtful and I just increased my subjective probability that this 3+-sigma result is wrong as well to 90 percent.

## No comments:

## Post a Comment