---------------------------------
Readme swiftstat
---------------------------------
D.J. Schad, Nov 2012
---------------------------------

During the simulations, the swiftstat displays some information in the console:

- 'Simulation parameters' shows the experiment that is simulated, and the number of runs of the model per subject. (settings from "swiftstat.inp")
- '1. Model parameter ranges' indicates the model parameters that were optimized during model fitting, and the ranges of valid parameter values.
- '2. SWIFT simulation parameters' shows best-fitting parameter values from the genetic algorithm-based fitting procedure. (sd.psc.par, sd.shuf.par)
- '3. Saccade error parameter': parameters defining random (noise) and systematic (SRE) error in saccade targeting for different types of saccades; these can be changed manually by editing the files sacc.par (for normal text) and sacc.shuf.par (for shuffled text)
- '3. Empirical data': dmin, dstep, and dbin provide information on binning of the distribution of SFD for computin fit

Goodness-of-fit statistics:
- dev[..] = deviance measure for different eye movement statistics (for details see compare4.c)

- dev[1] = first of multiple fixation durations
- dev[2] = second fixation durations
- dev[3] = single fixation durations
- dev[4] = total reading time
- dev[5] = skipping probability
- dev[6] = 2-fixation probability
- dev[7] = three or more fixations probability
- dev[8] = regression count (per regression target)

- dev[T] = deviation of fixation duration measures (dev[1]+dev[2]+dev[3]+dev[4])
- dev[P] = deviation of fixation probability measures (dev[5]+dev[6]+dev[7]+dev[8])
- dev[D] = deviation of the distribution of SFD (not included in optimization)
- dev[L] = deviation of distributed processing effects (lag, immediacy, successor effects)

- dev = total deviance ( dev[T]+dev[P]+dev[L] )
