G. Norris
Collecting karmic data since February 2026.
3. Observation Log
| Type |
Description |
Category |
K |
Date |
📭
Nothing recorded yet. Your karma ledger is blank.
Introduction
The concept of karma - that one's actions generate consequences of corresponding moral valence - appears across cultures, religions, and folk wisdom. Yet no rigorous attempt has been made to measure it. We propose a crowdsourced observational study.
Definition 1 (Actions).
An
action $a_i$ is a deed performed by the subject, characterized as morally positive or negative, and quantified by intent, effort, and magnitude.
Definition 2 (Events).
An
event $e_j$ is something that
happened to the subject, characterized by severity and importance.
The Karmic Input Function
For each action $a_i$ reported by a subject, we compute a karmic input value:
$$K_I(a_i) = I_i \cdot E_i \cdot M_i \cdot F_b$$ (1)
| Variable | Name | Domain | Description |
|---|
| $I$ | Intent | $[-1, +1]$ | $+1$ = saintly; $0$ = accident; $-1$ = evil, premeditated |
| $E$ | Effort | $[0, 1]$ | $0$ = "I didn't do anything"; $1$ = "with religious fervor" |
| $M$ | Magnitude | $(0, \infty)$ | Lives affected. Log scale: $10^{-3}$ (some grass) to $10^{10}$ (all humanity) |
| $F_b$ | Feedback | $[0, \infty)$ | Personal amplifier. See section 4. |
Constraint: if you were conscious of your action, then $M \geq 1$.
Magnitude mapping
The magnitude scale is logarithmic. We map a slider $s \in [0, 100]$ to:
$$M = 10^{(s/100) \cdot 10.3 - 3}$$ (2)
The Karmic Output Function
$$K_O(e_j) = S_j \cdot P_j$$ (3)
| Variable | Name | Domain | Description |
|---|
| $S$ | Severity | $[-1, +1]$ | $+1$ = life-changing good; $-1$ = devastating. Guilt + pride are clues. |
| $P$ | Importance | $[0, 1]$ | $0$ = forgettable; $1$ = defining moment |
The Feedback Factor
Definition 3 (Feedback Factor).
$$F_b^{(k)} = \varepsilon_k \cdot \iota_k$$
where $\varepsilon \in [0,1]$ is
ego (self-awareness) and $\iota \in [0,1]$ is
influence (social reach).
| $\varepsilon$ | Ego |
|---|
| 0 | Robot |
| 0.1 | "I am alive" |
| 0.2 | "I am a plant" |
| 0.3 | "I recognize myself" |
| 0.5 | "I recognize my future" |
| 0.7 | "I am more important" |
| 0.9 | "I am so f***ing hot!" |
| 1.0 | "I am a jealous god" |
| $\iota$ | Influence |
|---|
| 0 | Not even self |
| 0.1 | Vegan hermit |
| 0.3 | Farmer |
| 0.5 | Blue collar |
| 0.7 | Entertainer |
| 0.9 | Head of state |
| 1.0 | CEO of multinational |
| $\star$ | God |
Observation 1.
When $F_b = 1$ for all subjects (robot world), cause maps to effect without distortion.
We're all robots.
Observation 2.
As $F_b \to \infty$ (jealous god limit), even infinitesimal actions produce infinite karmic input.
We'd all be jealous.
Influence is not magnitude. Magnitude counts those affected at once; influence is the persistent structural capacity to multiply effects over time.
The Karma Ratio
$$K^{(k)} = \frac{\displaystyle\sum_{j=1}^{m} K_O(e_j)}{\displaystyle\sum_{i=1}^{n} K_I(a_i)} = \frac{\displaystyle\sum_{j=1}^{m} S_j \cdot P_j}{\displaystyle\sum_{i=1}^{n} I_i \cdot E_i \cdot M_i \cdot F_b^{(k)}}$$ (4)
The Karma Conjecture.
If karma exists as a natural phenomenon, then for sufficiently large population $N$ over sufficient time $T$:
$$\lim_{N \to \infty,\; T \to \infty} \;\frac{1}{N}\sum_{k=1}^{N} K^{(k)} \;\to\; 1$$
| Outcome | Interpretation |
|---|
| $\bar{K} \to 1$ | Karma is real (or behaves as if it is) |
| $\bar{K} \to 0$ | Anti-karma: the universe is indifferent |
| $\bar{K} < 0$ | Inverse karma: good deeds attract misfortune |
| $\bar{K} \gg 1$ | The universe is generous |
| $\text{Var}(K) \to 0$ | Deterministic karma |
| $\text{Var}(K) \gg 0$ | Stochastic karma |
Limitations and Selection Bias
The primary threat to validity is selection bias. Happy people may over-report good events; guilty people may over-report bad deeds.
(a) Volume. With enough participants, individual biases average out.
(b) Within-subject analysis. Does each person's $K$ converge over time?
(c) Category decomposition. Does karma operate locally - health karma, financial karma?
(d) Temporal analysis. If karma has a time constant $\tau$, we should see $K \to 1$ only over windows $\gg \tau$.
If the bias is the finding - "optimists experience more good events," that's interesting too. It either proves karma or reveals something about human self-perception. Either result is worth having.
Future Extensions
Population-level statistics. Histogram of $K$ across the user base. Is it normal around 1? Bimodal?
The karma weather. Real-time aggregate $\bar{K}_{\text{today}}$ - is humanity trending good or bad?
Time-series correlation. Cross-correlation between $K_I$ and $K_O$ time series to find the "speed of karma."
Entropy analysis. Karmic entropy $H(K) = -\sum p(K) \log p(K)$. Low entropy = lawlike; high = noise.
Acknowledgments. This framework was first sketched on paper in November 2000 -
17 pages of handwritten notes, formulas, and UI wireframes, filed under
C:\Documents and Settings\glen\My Documents\websites\karmal.... Twenty-five years later, the web caught up.
Add art to science + pop culture. It could be beautiful. If it ends up doing a beautiful thing, you'd want it to be.
"Construct the form in such a way that it is an exercise which encourages one to think in terms of maximizing their positive karma. Same world."