Update FLOWERY_NONSENSE_PERHAPS.md

Author: microprediction

FLOWERY_NONSENSE_PERHAPS.md

@@ -1,62 +1,62 @@
-# Using Markowitz‐Style Optimization for Bouquet Design
+# Using Markowitz-Style Optimization for Bouquet Design
 
 ## Can We Apply Markowitz Directly?
 
-> **Short answer:** Yes, you **can** cast bouquet design as a mean–variance problem, but it takes extra effort because beauty is not a simple, linear “return,” and aesthetic “risk” is tricky to define.
+> **Short answer:** Yes, you *can* cast bouquet design as a mean–variance problem, but it takes extra effort because beauty is not a simple, linear “return,” and aesthetic “risk” is tricky to define.
 
 ---
 
-## 1  Mapping Markowitz Terms to Floral Design
+## 1 Mapping Markowitz Terms to Floral Design
 
-| Markowitz Term | Floral Analogue | How You Might Measure It |
-| -------------- | --------------- | ------------------------ |
-| **Asset weight \(w_i\)** | Proportion or count of stems of flower *i* | Decision variables (often integers) |
-| **Expected return \(\mu_i\)** | Average beauty contribution of one stem of flower *i* | Crowd-sourced ratings, expert scores, ML predictions |
-| **Covariance \(\Sigma_{ij}\)** | Disharmony/clash between flowers *i* and *j* | Pair-wise harmony scores, color-wheel distance, shape-clash metrics |
-| **Portfolio return \(w^\top\mu\)** | Total predicted beauty | Linear sum of contributions |
-| **Portfolio risk \(w^\top\Sigma w\)** | Aggregate discordance risk (lack of unity) | Quadratic clash penalty |
+| Markowitz term                     | Floral analogue                                             | How you might measure it |
+|-----------------------------------|-------------------------------------------------------------|--------------------------|
+| **Asset weight $w_i$**            | Proportion or count of stems of flower _i_                  | Decision variables (often integers) |
+| **Expected return $\mu_i$**       | Average beauty contribution of one stem of flower _i_       | Crowd-sourced ratings, expert scores, ML predictions |
+| **Covariance $\Sigma_{ij}$**      | Disharmony/clash between flowers _i_ and _j_               | Pairwise harmony scores, colour-wheel distance, shape-clash metrics |
+| **Portfolio return $w^\top\mu$**  | Total predicted beauty                                      | Linear sum of contributions |
+| **Portfolio risk $w^\top\Sigma w$** | Aggregate discordance risk (lack of unity)                 | Quadratic clash penalty |
 
 **Continuous formulation**
 
-\[
+$$
 \begin{aligned}
-\max_{w}\;& w^\top\mu \;-\;\lambda\,w^\top\Sigma w \\[6pt]
+\max_{w}\;& w^\top\mu \;-\;\lambda\,w^\top\Sigma w \\[4pt]
 \text{s.t. }& \mathbf 1^\top w = 1,\quad w \ge 0
 \end{aligned}
-\]
+$$
 
 ---
 
-## 2  Why Most Florists Don’t Stop Here
+## 2 Why Most Florists Don’t Stop Here
 
-| Issue | Why It Matters for Bouquets |
-| ----- | --------------------------- |
-| **Non-additive beauty** | Synergies/clashes are often *non-linear*. |
+| Issue | Why it matters for bouquets |
+|-------|-----------------------------|
+| **Non-additive beauty** | Synergies and clashes are often *non-linear*. |
 | **Risk ≠ variance** | A symmetric variance treats all deviations equally; humans don’t. |
-| **Tiny integer portfolios** | 12–24 stems → integer quadratic program, not the smooth QP Markowitz likes. |
-| **Data burden** | Need an *N × N* clash matrix; expensive to collect. |
-| **Beauty is multi-objective** | People like *unity and variety* simultaneously (an inverted-U curve). |
+| **Tiny integer portfolios** | 12–24 stems → an integer quadratic program, not the smooth QP Markowitz likes. |
+| **Data burden** | Need an $N \times N$ clash matrix; expensive to collect. |
+| **Beauty is multi-objective** | People like *unity **and** variety* simultaneously (an inverted-U curve). |
 
 ---
 
-## 3  Making Markowitz Useful Anyway
+## 3 Making Markowitz Useful Anyway
 
-1. **Rate single stems** → get \(\mu_i\).  
-2. **Rate pairs of stems** → build clash matrix \(\Sigma\).  
+1. **Rate single stems** → get $\mu_i$.  
+2. **Rate pairs of stems** → build clash matrix $\Sigma$.  
 3. **Add florist constraints**  
-   - Integer counts  
-   - Budget and stem limits  
-   - Colour quotas, minimum greens, etc.  
+   * Integer counts  
+   * Budget and stem limits  
+   * Colour quotas, minimum greens, etc.  
 4. **Solve the integer QP** (branch-and-bound or meta-heuristic).  
 5. **Designer tweak pass** to fine-tune beyond the model.
 
 ---
 
-## 4  Alternatives & Hybrids
+## 4 Alternatives & Hybrids
 
-* **Multi-objective genetic / flower-pollination algorithms**: naturally handle unity *and* variety.  
-* **Black–Litterman-style priors**: blend expert views with crowd data.  
-* **Robust or CVaR risk measures**: penalise ugly clashes more than mild disharmony.
+* **Multi-objective genetic / flower-pollination algorithms** — naturally handle unity *and* variety.  
+* **Black–Litterman-style priors** — blend expert views with crowd data.  
+* **Robust or CVaR risk measures** — penalise ugly clashes more than mild disharmony.
 
 ---