PAIesque
将心率数据转化为训练负荷和恢复指标
新版本 82
# Release Notes – v82
**Sharing & Device Names**
- Shared recordings now include the original device display name directly in the file, ensuring correct names (e.g. "HRM 600" instead of "HRM_600") appear in group member lists and the device spinner.
- The server correctly reads device names from the new CSV column, and accepts arbitrary BLE advertising names via URL encoding.
- Status updates now only upload for actively recording devices, preventing stale data from idle sensors.
**Map & Remot
**Sharing & Device Names**
- Shared recordings now include the original device display name directly in the file, ensuring correct names (e.g. "HRM 600" instead of "HRM_600") appear in group member lists and the device spinner.
- The server correctly reads device names from the new CSV column, and accepts arbitrary BLE advertising names via URL encoding.
- Status updates now only upload for actively recording devices, preventing stale data from idle sensors.
**Map & Remot
PAIesque helps athletes and fitness enthusiasts monitor their training through a simple three-step logic:
1. Measure training impulse (TRIMP) — The faster and longer your heart beats, the higher your daily score.
2. Analyze patterns over time — Track how your TRIMP accumulates and distributes
3. Monitor your body's response — Compare how your body reacts to training load
PAIesque is different from commercial fitness apps (e.g. Garmin,
Whoop, Polar). Every metric comes from published, peer-reviewed research with transparent methods that can be calculated from heart rate data alone. The app only includes metrics we can verify and reproduce from first principles — no proprietary black boxes, no undisclosed algorithms. And, all your data stays on your device.
1. TRIMP:
• Banister TRIMP — The original exponential model with sex-specific coefficients (a=0.64/0.86, b=1.92/1.67) [Banister, 1991; Morton et al., 1990]
• iTRIMP — Individualized TRIMP with customizable b coefficient (1.5-4.0) [Stagno et al., 2007; Akubat et al., 2012]
• LT-TRIMP — Lactate Threshold-based model with β coefficient (0.04-0.11) and smooth transition at LT [Cheng et al., 1992; Mader et al., 1976; Gaesser and Poole, 1986]
• PAI-esque — PAI-inspired metric using EWMA (not the official commercial algorithm) [Nes et al., 2017; Kieffer et al., 2021]
2. Patterns over time:
• Intensity zones — Time and TRIMP spent in low/moderate/high zones (polarized training model) [Seiler and Tønnessen, 2009; Stöggl and Sperlich, 2014]
• EWMA — Exponentially Weighted Moving Average for rolling loads (more sensitive than simple averages)
• ACWR — Acute:Chronic Workload Ratio for injury risk monitoring (0.8-1.3 = sweet spot) [Murray et al., 2017; Griffin et al., 2021; Gabbett, 2016]
• Polarized Training Score — 0-100 measure of how closely your distribution matches your targets
3. Body's response:
• Resting Heart Rate (RHR) — Calculated from your defined sleep window (adaptive percentile: 5th-15th)
• Heart Rate Variability (HRV) — Daily RMSSD averages during sleep [Task Force, 1996; Plews et al., 2013; Buchheit, 2014]
• EWMA trends — Exponentially weighted moving averages for both RHR and HRV (acute and chronic windows)
• Combined interpretation — RHR ↓ + HRV ↑ = positive adaptation; RHR ↑ + HRV ↓ = possible fatigue
Data Management:
• CSV export/import
• Complete backup/restore (db)
• All data stays on your device — no accounts, no cloud uploads, no tracking
Creative Use Cases:
• Coach analyzing athletes — Import athlete exports, analyze charts, provide feedback
• Research analysis — Export CSV files for custom analysis in R, Python, or spreadsheets
• Switch between athletes — Use "Delete All Data" + CSV import to analyze multiple individuals
Requirements:
• Google Health Connect installed on your device
• Heart rate (and HRV) data in Health Connect from your wearable device (Gadgetbridge, Garmin, Polar, Samsung, etc.)
• Android 8.0 (API 26) or higher
Note on PAI:
Our PAI-esque implementation is NOT the official commercial PAI® algorithm (which is proprietary). It uses EWMA and scaled TRIMP values to provide a similar intensity-weighted weekly score. The 100 PAI target remains the evidence-based health outcome from the HUNT Study research.
1. Measure training impulse (TRIMP) — The faster and longer your heart beats, the higher your daily score.
2. Analyze patterns over time — Track how your TRIMP accumulates and distributes
3. Monitor your body's response — Compare how your body reacts to training load
PAIesque is different from commercial fitness apps (e.g. Garmin,
Whoop, Polar). Every metric comes from published, peer-reviewed research with transparent methods that can be calculated from heart rate data alone. The app only includes metrics we can verify and reproduce from first principles — no proprietary black boxes, no undisclosed algorithms. And, all your data stays on your device.
1. TRIMP:
• Banister TRIMP — The original exponential model with sex-specific coefficients (a=0.64/0.86, b=1.92/1.67) [Banister, 1991; Morton et al., 1990]
• iTRIMP — Individualized TRIMP with customizable b coefficient (1.5-4.0) [Stagno et al., 2007; Akubat et al., 2012]
• LT-TRIMP — Lactate Threshold-based model with β coefficient (0.04-0.11) and smooth transition at LT [Cheng et al., 1992; Mader et al., 1976; Gaesser and Poole, 1986]
• PAI-esque — PAI-inspired metric using EWMA (not the official commercial algorithm) [Nes et al., 2017; Kieffer et al., 2021]
2. Patterns over time:
• Intensity zones — Time and TRIMP spent in low/moderate/high zones (polarized training model) [Seiler and Tønnessen, 2009; Stöggl and Sperlich, 2014]
• EWMA — Exponentially Weighted Moving Average for rolling loads (more sensitive than simple averages)
• ACWR — Acute:Chronic Workload Ratio for injury risk monitoring (0.8-1.3 = sweet spot) [Murray et al., 2017; Griffin et al., 2021; Gabbett, 2016]
• Polarized Training Score — 0-100 measure of how closely your distribution matches your targets
3. Body's response:
• Resting Heart Rate (RHR) — Calculated from your defined sleep window (adaptive percentile: 5th-15th)
• Heart Rate Variability (HRV) — Daily RMSSD averages during sleep [Task Force, 1996; Plews et al., 2013; Buchheit, 2014]
• EWMA trends — Exponentially weighted moving averages for both RHR and HRV (acute and chronic windows)
• Combined interpretation — RHR ↓ + HRV ↑ = positive adaptation; RHR ↑ + HRV ↓ = possible fatigue
Data Management:
• CSV export/import
• Complete backup/restore (db)
• All data stays on your device — no accounts, no cloud uploads, no tracking
Creative Use Cases:
• Coach analyzing athletes — Import athlete exports, analyze charts, provide feedback
• Research analysis — Export CSV files for custom analysis in R, Python, or spreadsheets
• Switch between athletes — Use "Delete All Data" + CSV import to analyze multiple individuals
Requirements:
• Google Health Connect installed on your device
• Heart rate (and HRV) data in Health Connect from your wearable device (Gadgetbridge, Garmin, Polar, Samsung, etc.)
• Android 8.0 (API 26) or higher
Note on PAI:
Our PAI-esque implementation is NOT the official commercial PAI® algorithm (which is proprietary). It uses EWMA and scaled TRIMP values to provide a similar intensity-weighted weekly score. The 100 PAI target remains the evidence-based health outcome from the HUNT Study research.
版本
尽管下面提供了 APK 安装包的下载选项，但你应该注意，以这种方式安装将不会收到更新通知，这是一种不太安全的下载方式。 我们建议你安装使用 F-Droid 客户端。
下载 F-Droid-
arm64-v8aarmeabi-v7ax86x86_64该版本需要 Android 14 及以上版本。
此包由原始开发者构建并签名，并保证对应于此源代码 tarball。
-
arm64-v8aarmeabi-v7ax86x86_64该版本需要 Android 14 及以上版本。
此包由原始开发者构建并签名，并保证对应于此源代码 tarball。
-
arm64-v8aarmeabi-v7ax86x86_64该版本需要 Android 14 及以上版本。
此包由原始开发者构建并签名，并保证对应于此源代码 tarball。
新版本 81
# Release Notes – v81
## Critical Bug Fix
**Empty device spinner after first recording**
Previously, after recording a session (with or without a BLE sensor), the device spinner in the analysis screen remained empty, preventing users from viewing their training data. This release completely decouples the recording feature from `paiesque.db` and ensures that all recorded devices appear immediately in the analysis view.
## New Features & Improvements
- **Removed misleading toast** – The “No sa
## Critical Bug Fix
**Empty device spinner after first recording**
Previously, after recording a session (with or without a BLE sensor), the device spinner in the analysis screen remained empty, preventing users from viewing their training data. This release completely decouples the recording feature from `paiesque.db` and ensures that all recorded devices appear immediately in the analysis view.
## New Features & Improvements
- **Removed misleading toast** – The “No sa
新版本 80
# Release Notes – v80
## New Features
### Unified Navigation Bar
The toolbar now contains a **three‑button toggle** (Trends / Record / Analysis) that highlights the current activity. The Record button **turns red** while a recording is in progress.
### Distance Calculation Settings
- Configurable **distance step size** (raw … 100 m) and **minimum time interval** (none … 30 s).
- Access via recording screen menu → *Distance Calculation*.
- Applies to live recording and historical analysis.
##
## New Features
### Unified Navigation Bar
The toolbar now contains a **three‑button toggle** (Trends / Record / Analysis) that highlights the current activity. The Record button **turns red** while a recording is in progress.
### Distance Calculation Settings
- Configurable **distance step size** (raw … 100 m) and **minimum time interval** (none … 30 s).
- Access via recording screen menu → *Distance Calculation*.
- Applies to live recording and historical analysis.
##