# Changes: Discrete Demand & Tiered Historical Display

## Summary
Made demand values discrete (integer-only) and split historical demand visibility into two tiers:
- **Basic (everyone)**: See past discrete demand values only
- **Advanced (forecasting tool users only)**: See past demand values + summary statistics (mean, std, min, max, n)

## Changes Made

### 1. Discrete Demand Generation (`newsvendor/__init__.py`)

#### In `creating_session()`:
- **Line ~90**: Synthetic history generation now rounds to integers:
  ```python
  hist = [float(max(0, round(np.random.normal(mean_hist, std_hist)))) for _ in range(C.HISTORY_SAMPLES)]
  ```

- **Line ~110**: Fallback synthetic history also rounds to integers:
  ```python
  hist = [float(max(0, round(np.random.normal(mean_hist, std_hist)))) for _ in range(C.HISTORY_SAMPLES)]
  ```

- **Line ~120**: Actual demand per round now rounded to integers:
  ```python
  draw = np.random.normal(player.mean_demand, player.std_demand)
  player.actual_demand = float(max(0, round(draw)))
  ```

### 2. Historical Demand Display (`Order.vars_for_template`)

#### Formatting changes:
- **Line ~177**: Format history as integers without decimals:
  ```python
  formatted_history = [f"{int(v)}" for v in actual_hist]
  ```

- **Line ~180-181**: Min/max stored as integers:
  ```python
  hist_min = int(np.min(actual_hist)) if actual_hist else None
  hist_max = int(np.max(actual_hist)) if actual_hist else None
  ```

- **Line ~338-339**: Min/max passed to template as integers:
  ```python
  min=hist_min if hist_min is not None else None,  # Already int
  max=hist_max if hist_max is not None else None,  # Already int
  ```

### 3. Results Page Display

#### In `Results.vars_for_template`:
- **Line ~415**: Actual demand shown as integer:
  ```python
  actual_demand=int(player.actual_demand),  # Display as discrete integer
  ```

#### In feedback display:
- **Line ~314**: Last round's demand shown as integer:
  ```python
  actual=int(pv['actual_demand_history'][idx]) if pv.get('actual_demand_history') else None,
  ```

### 4. Template Changes (`newsvendor/Order.html`)

#### Historical Demand section (Lines 17-32):
- Everyone sees: List of discrete past demand values
- Forecasting tool users additionally see: Summary statistics in a separate line below the values

**Before:**
```html
<p style="margin:4px 0 8px;">Past actual demand from your completed rounds.</p>
<p style="margin:0 0 6px;">
    mean={{ history_stats.mean }}, std={{ history_stats.std }}, ...
</p>
<div><!-- demand values --></div>
```

**After:**
```html
<p style="margin:4px 0 8px;">Past actual demand from your completed rounds:</p>
<div><!-- demand values --></div>
{% if show_demand_forecast %}
<p style="margin:10px 0 0;">
    <strong>Summary statistics:</strong> mean={{ history_stats.mean }}, ...
</p>
{% endif %}
```

## Access Control by Tool Level

| Tool Count | Access Level | What They See |
|-----------|--------------|---------------|
| 1 (None) | Basic | Past demand values only (after Round 1) |
| 2+ (Forecasting) | Advanced | Past demand values + summary stats (mean, std, min, max, n) |

## Testing Recommendations

1. **Test discrete demand generation**:
   - Start a new session
   - Check that `actual_demand` values are always integers (no decimals)
   - Verify on Results page

2. **Test basic access (tool_count=1)**:
   - Assign participant with `tool_count=1`
   - After Round 1, verify Historical Demand shows only discrete values
   - Confirm NO summary statistics are visible

3. **Test advanced access (tool_count>=2)**:
   - Assign participant with `tool_count=2` or higher
   - After Round 1, verify Historical Demand shows discrete values AND summary stats
   - Confirm stats line appears below the demand values

4. **Test data consistency**:
   - Verify demand values match between:
     - Results page display
     - Historical Demand on next Order page
     - Feedback tool (if available)
     - Dashboard charts (if available)