It’s been over a month since my last post, and a lot has shifted for me.
When I built out Experiments 1 and 2, I thought the major challenges in validating psi research lay in experimental setup. I believed experimenters choosing which parts of data to analyze, dropping data points, data not being encrypted to dissuade accusations of alteration, collecting too few samples, or influencing results simply by being present in the room had added unnecessary noise and doubt. There were also missing controls, like a parallel “ghost” stream to test alongside human data. So I built 1 and 2 to remove those issues entirely.
I was in such a rush, though. I only have six months before I have to return to working for someone else, and that deadline meant I didn’t spend time digging into prior research. I had skimmed the PEAR Lab studies and the Global Consciousness Project and the usual criticisms, but I had not spent time really digging into how much rigorous work already been done. Once I finally took the time to dive in, I am coming back feeling incredibly humbled.
There’s an extraordinary amount of serious, rigorous research in psi. Many studies were meticulously controlled, replicated, double-blinded, and statistically sound in ways that are hard to reproduce outside a lab. The effects are small, but they’re consistent enough to suggest something is there. That convinced me that psi, if real (which I think it is after reading through various studies), operates at a very low level. Running yet another “does psi exist?” experiment now feels redundant. The bar is already high, and I doubt I could surpass it, or even match it. More importantly, I don’t think that’s where the interesting questions are anymore.
A Shift in Focus
What caught my attention recently was a passing line in Maier and Dechamps, Intentional Observer Effects on Quantum Randomness: A Bayesian Analysis Reveals Evidence Against Micro-Psychokinesis (2018) that described a dampening harmonic oscillation over time in their post analysis. Although their Bayesian analysis pointed toward no psychokinetic effect*, their follow-up analysis showed a rhythmic decline in effect strength found across participants but didn’t appear in automated controls. That phrasing caught me, because it suggests a physical signature that could be tracked rather than trying to persuade people with a simple statistical anomaly for an extremely small effect.
That idea changed my trajectory a bit. If psi operates through information or entropy, maybe it leaves systematic traces in the structure of randomness itself. The fact that there was a recognizable temporal shape fascinated me. It suggests that if psi has any real influence, it might leave traces as patterns of information across time.
*(Their Bayesian result doesn’t make me think psi is impossible. They assumed a moderate effect size, while most psi research shows effects that are much smaller. Under that assumption, their analysis was really testing whether a moderate psi effect exists, and it found none. That makes sense, especially with more than twelve thousand online participants working on their own computers, with no control over environment, focus, or selection. I see their conclusion as valid for what they tested: there is no medium-sized effect in a wide, uncontrolled population.)
What interested me most in their work is not the null result but the pattern. The idea that intention might have temporal dynamics is what inspired Experiment 3.
Related Work on Temporal Structure in PSI
Oscillatory or declining effects in micro-PK
Maier and colleagues found that micro-PK data may follow a damped, oscillatory decline when compared to simulated controls. My work takes that idea forward but measures it directly within each session using a real-time ghost control and a bit-indexed timeline.
Time-series reanalysis of RNG databases
Varvoglis and Bancel re-examined PEAR Lab data using time-series methods and showed that psi effects might fluctuate dynamically rather than remaining constant. My entropy and autocorrelation analyses build on this kind of process-based approach.
Anticipatory timing in physiology
Meta-analyses of pre-stimulus physiological responses (e.g., Mossbridge et al.) show small but consistent effects seconds before unpredictable events. While I’m not studying physiology, this work supports the idea that psi, if it exists, might express itself through timing patterns which I have included in both Experiment 2 & 3.
Why Experiments 3
Experiments 1 and 2 will keep running on autopilot until they reach three hundred participants, at which point I’ll analyze them properly. My active focus is now on Experiments 3 (and 4 coming soon!), which are designed to look for structure, and not just statistical anomalies.
Experiment 3 examines raw signatures and searches for structure in randomness. For every quantum bit used in the intention task, I now pull the full byte it came from. That allows me to look for correlations, repeating sequences, or subtle deviations that might reveal hidden order inside what should be pure randomness. It uses A LOT of resources from the QRNG so you may have to schedule with us to take it if you are receiving messages that we have run out of data.
I’m also exploring whether exposure to high-entropy versus low-entropy feedback affects how easily participants can interact or “tune in.” If consciousness interacts with entropy, the amount of visible order might change how the effect expresses itself.
The Core Design
- Subject stream: the data participants attempt to influence, shown visually in real time.
- Demon stream: an identical, yoked control feed that runs in parallel but is never displayed.
The demon stream acts as a silent twin. If both streams behave the same, the system is clean. If they diverge, something interesting is happening.
Key improvements include:
- A live demon stream that controls for shared technical artifacts.
- Bit-indexed timing, which removes the influence of variable rest pauses.
- Two separate measures of dependence: autocorrelation in data bits (physical structure) and autocorrelation in performance hits (psychological reinforcement).
The Feedback Environment: Ring vs Mosaic
Participants alternate between two feedback styles:
- Ring mode: ordered, rhythmic visuals that feel coherent.
- Mosaic mode: chaotic, high-entropy visuals with little obvious structure.
Both draw from the same quantum data. What changes is the way that data is represented. This pilot doesn’t start with a firm prediction. Low-entropy feedback might make focusing easier, but high-entropy feedback could offer more room for subtle effects to appear. For now, the goal is simply to see whether the visual form of information matters.
What I’m Measuring
- Entropy suppression
Every 1,000 bits, I compute Shannon entropy. If intention creates local order, the subject stream might drift slightly below the ghost. - Autocorrelation
Within each 30-second block, I measure how each outcome depends on the previous one. The code computes autocorrelation for both performance hits and raw bits. - Damping over time
If psi behaves like a resonance, early structure might fade as entropy restores itself. I track this by bit index rather than clock time because participants take breaks at different moments.
The CIR²S Framework
All of this fits into my broader model:
Coherence → Information → Resonance → Redundancy → S-Selection.
Intention may create temporary coherence (entropy suppression). Feedback can amplify it (resonance and redundancy). Visual form influences how the effect appears (S-selection). Finally, thermodynamic processes restore equilibrium, producing the damping pattern I’m testing for.
What This Experiment Is (and Isn’t)
This is a pilot, an exploratory study of how information behaves when consciousness interacts with quantum randomness. The goal is discovery.
If the subject and ghost streams are indistinguishable, then this design doesn’t reveal anything new. If subtle patterns appear then I’ll know where to focus next. Either result is informative.
Closing Thoughts
Earlier generations of psi research have already done careful scientific work. My shift is just one of emphasis. Instead of asking whether psi exists, I’m asking how it might behave if it does.
The effect, if real, seems tiny. By looking at structure instead of summary statistics, we may find patterns that would otherwise be invisible. Looking for anomalies in randomness is like listening to an orchestra and waiting for a single wrong note.
It’s rare, fleeting, and easy to explain away as noise.
Looking for temporal structure is like stepping back and listening for a faint melody woven through the entire performance. A single wrong note tells you almost nothing, but a repeating phrase, even if soft, tells you the orchestra is following some underlying pattern.
Experiment 3 is my attempt to explore those patterns and to trace whatever signatures might be hiding out in the noise. Because signatures are harder to write off than statistical anomalies.
Here is a link to Experiment 3 if you are interested in contributing some data points!
Acknowledgements
What The Quark LLC would like to thank ChatGPT and Claude for help creating this blog post. This experiment runs on quantum random numbers from Cisco Outshift and LFDR. Special thanks to both teams for maintaining these critical scientific resources and making them freely available to researchers.
Searching For Structure In Randomness © 2025 by What The Quark LLC is licensed under Creative Commons Attribution-NonCommercial 4.0 International
You are free to share, copy, and adapt the material for noncommercial purposes, provided that appropriate credit is given.
Leave a Reply