To further evaluate the bunch fitting results (https://drupal.star.bnl.gov/STAR/blog/dilks/relative-scaler-asymmetries), I've started searching for any correlations between spin patterns and the fit results.

ASYMMETRY DEFINITIONS
BUNCH FITTING ALGORITHM

Everything below is for Run13 only

-----[ Run 13 Spin Patterns]-----------------------------------------------

In Run13, for both the blue and yellow beams, we had 4 spin patterns, which repeat modulo 8 bXings. The spin pattern starts at bunch=0. I've assigned a beam pattern number as follows:

Let T_B and T_Y denote the blue beam and yellow beam pattern numbers, respectively.
The STAR bXing pattern number, T_bx, is defined as:

T_bx = 10 T_B + T_Y

In Run13, only beam patterns {1,2} were collided with {3,4} and vice versa, leading to only 8 possible bXing pattern numbers:

{13, 14, 23, 24, 31, 32, 41, 42}

Noting that the two abort gaps were 9 bXings each, there are nominally 102 bXings with filled bunches in every fill (sometimes there are empty / deviant bunches, but let us only consider nominal fills where all 102 possible bXings are filled in both beams)

We can count the number of bXings for each of the four proton spin states {- -, - +, + -, + +} for each of the 8 bXing pattern numbers:

For a nominally filled fill, it is always the case that three of the spin states occur in 26 bXings each, and one spin state occurs in 24 bXings. In the table above, I've made the number "24" bold to make it easy to see which spin state is the outlying spin state for each bXing pattern number.

Based on these spin pattern definitions, I've added colours to the ε vs. run index plots from the previous blog entry. The colours denote the spin pattern. I chose the same colour for the case if we switch the blue and yellow beams (e.g. bXing pattern 13 is coloured the same as 31):

OUTLYING SPIN STATES:

• For RED and GREEN, the outlying spin state is either (+ -) or (- +)  [patterns 13,31 and 24,42]
• For MAGENTA and BLUE, the outlying spin state is either (+ +) or (- -) [patterns 14,41 and 23,32]

-----[ Pattern-Coloured ZDC Singles Asymmetries w.r.t. VPDX ]-----------------------------------------------

The plots below are the coloured versions of the asymmetry plots shown on the previous blog entry; the fit lines are through all points & spin patterns

Observations:

• Asymmetry 3 (double spin asymmetry) is correlated with spin pattern
• Asymmetries 5,6,7, and 9 (the "oddball" false asymmetries) are also correlated with spin pattern
• I tested the sensitivity of the correlation in the following ways:
  • "spin state equalisation" -- randomly remove the minimum number of bXings such that the number of valid bXings considered for each spin set was equal
    • Little to no effect on the correlation... maybe a coding error somewhere???
  • Only consider a set of 24 bXings (72-103) so that if there are no deviant bXings in that range, the number of bXings for each spin state should be equal
    • (note also that 72 = 0 (mod 8), i.e., bXing 72 has same spin as bXing 0)
    • Statistics are worse, but the correlation is still visible; scroll down to the "EXTRA PLOTS" section to see it
    • Maybe it's not a coding error after all...

ZDC EAST RAW ASYMMETRIES

ZDC WEST RAW ASYMMETRIES

-----[ EXTRA PLOTS ]-----------------------------------------------

Here are the asymmetries where we consider only bXings 72-103 (see "Observations" in above section for details)

Additional observation: bXings 69 and 70 were empty for the first ~230 runs in this set of runs considered. The bXing range that is selected does not include these two bXings, yet we still see an enhancement of the spin pattern correlation for the first ~230 runs!

ZDC EAST RAW ASYMMETRIES FOR BXINGS 72-103

ZDC WEST RAW ASYMMETRIES FOR BXINGS 72-103