Monitoring Timing in MMI experiments

So according to Scott’s papers the the average drift velocity or how fast the walker moves towards a bound is equal to the effect size of mental influence.

In the basic experiment outlined in the introduction of scotts paper “Detecting the Presence of Anomalous Effects by Monitoring Timing in Random Walks” you measure the bias produced when an operator intends to produce more 0s or 1s, within 200ms to 1 second.

This is something I have independently verified. I have noticed while trying to replicate some of Scotts work, that when using the MED100Kx8, I noticed some bias, but only relatively quickly into the trial, within around 200ms. The average bias I noticed was also indeed around .5

I remember newton and 00 talking about the timing of the whole experiment years ago, has anyone else worked specifically with monitoring the timing? My initial intuition was that more randomness at a higher speed would always be better, but now after experimenting it seems that besides high entropy bandwidth, timing is extremely crucial. I notice a bias quickly, but it also trails off quickly (which may say something about the average operators ability to focus lol)

I guess I am just wondering if anyone else has focused on this element, and perhaps seeing if there is any more information about monitoring timing that scott has yet to put in a paper.

I am still working out how timing effects the experiment in free running control systems, vs synchronized or trial based systems. I am interested in free running systems, but it is much easier for me to understand systems that have user feedback to reinforce mental influence.

here is a basic outline of the experiment i have been working with

Control Trials:

Run the RNG without any intentional influence to measure the usual distribution of times taken to reach the bounds set within the RWBA algorithm.

Test Trials:

Each participant will engage in a series of trials where they will focus on mentally influencing the RNG to produce either more '1’s or more '0’s over a fixed time period, such as 200 ms, which aligns with the known trial timing for the device.

Random Walk Bound Selection:

Set a specific bound (eg n=31) for the RWBA algorithm to amplify the natural bias of the RNG output.

Recording and Time Measurement:

Record the time taken (number of steps) to reach the bound for each trial with the RNG.

Test Procedure:

Each test session will consist of alternating blocks of control and test trials, with the sequence randomized to prevent order effects.

Use the RWBA algorithm to process the output bits for each trial, tracking the number of steps taken to reach the bound.

Record the results from each participant, along with timestamps to analyze

Data Analysis:

Calculate the CDF from control trials and use it to assess the probability after each random walk output during test trials.

Use surprisal values derived from these probabilities as weights for the outcomes, comparing the weight distribution between control and test trials.

Apply statistical tests to compare the timing distributions between the control and test trials.

If mental influence is effective, the random walk should detect a significant deviation in timing during the test trials compared to the control trials, with statistical significance set at p < 0.05.

please let me know if this can be improved