# Data Quality

Vybe's commitment to data quality goes beyond vetted markets. Every endpoint is designed with accuracy as the priority—from holder data refreshed every 3 hours to trade history filtered for real transactions to PnL calculations that actually match reality.

## Why Data Quality Matters

On Solana, data quality varies wildly across providers. Common problems include:

| Problem                | Impact                         | Vybe Solution            |
| ---------------------- | ------------------------------ | ------------------------ |
| Stale holder data      | 24-48 hour delays              | Refreshed every 3 hours  |
| Fake price wicks       | False signals, triggered stops | Vetted markets only      |
| Wash trading in volume | Misleading activity            | Filtered out             |
| Inaccurate PnL         | Wrong profit calculations      | Vetted trade prices      |
| Unlabeled wallets      | Anonymous whale addresses      | 10,000+ labeled entities |
| Missing DeFi positions | Incomplete portfolio           | 50+ protocols tracked    |

⚠️ Bad data leads to bad decisions. That's why Vybe exists: to give developers and traders data they can rely on.

***

## How We Ensure Quality

| Data Type            | Quality Measures                                                                                       |
| -------------------- | ------------------------------------------------------------------------------------------------------ |
| **Token Prices**     | Aggregated from vetted markets only, weighted by liquidity depth, outliers filtered, real-time updates |
| **Holder Data**      | Refreshed every 3 hours, wallets labeled with known entities, categorized by holding size              |
| **Trade History**    | Only from vetted markets, wash trading filtered, full instruction metadata preserved                   |
| **PnL Calculations** | Uses vetted trade prices, accurate cost basis tracking, handles partial sells and transfers correctly  |
| **Candlestick Data** | Vetted pools only, no fake wicks, accurate OHLCV from real trades                                      |
| **Liquidity Data**   | Real-time depth from vetted pools, symmetric balance requirements                                      |

***

## Data Freshness

Different data types have different update frequencies based on what makes sense:

| Data Type        | Update Frequency   | Why                                   |
| ---------------- | ------------------ | ------------------------------------- |
| Token prices     | Real-time          | Trading decisions need current prices |
| Candles          | Per block (~400ms) | Charts should reflect latest trades   |
| Pool liquidity   | Real-time          | Depth changes with every trade        |
| Trade history    | Real-time          | Trades appear instantly               |
| Holder counts    | Every 3 hours      | Balance changes settle over time      |
| Top holders list | Every 3 hours      | Matches holder count updates          |
| PnL metrics      | On query           | Calculated fresh each request         |
| DeFi positions   | On query           | Fetched from protocols live           |

***

## Comparison: Vybe vs Typical APIs

| Metric                     | Vybe               | Typical Solana API |
| -------------------------- | ------------------ | ------------------ |
| **Price source**           | Vetted pools only  | All pools          |
| **Fake wick filtering**    | ✅ Automatic        | ❌ None             |
| **Wash trade filtering**   | ✅ Yes              | ❌ No               |
| **Holder refresh rate**    | 3 hours            | 24 hours           |
| **Holder labels**          | ✅ 10,000+ entities | ❌ None             |
| **DeFi position tracking** | ✅ 50+ protocols    | ❌ Not available    |
| **PnL accuracy**           | Vetted trades only | All trades         |
| **Cost basis tracking**    | ✅ Accurate         | ⚠️ Approximate     |

***

## Deep Dive: Holder Data Quality

Holder data is particularly important for token analytics. Here's how Vybe ensures accuracy:

### Refresh Frequency

* **Every 3 hours** vs industry standard of 24 hours
* Detect accumulation/distribution in near real-time
* Holder count metrics are actually useful

### Wallet Labels

* **10,000+ labeled entities** including:
  * CEX wallets (Binance, Coinbase, etc.)
  * KOLs and influencers
  * VC and fund wallets
  * Protocol treasuries
  * Market makers
* When a whale appears in top holders, you know who they are

### Size Categorization

Holders are categorized by position size:

* **Dust**: \< $1
* **Small**: $1 - $1,000
* **Medium**: $1,000 - $10,000
* **Large**: $10,000 - $100,000
* **Whale**: > $100,000

***

## Deep Dive: PnL Accuracy

Accurate PnL requires accurate trade data. Here's how Vybe gets it right:

### The Problem

```
User buys 1000 TOKEN at $0.01 (cost: $10)
Wash trader manipulates price to $0.10
Other API shows: +$90 profit (900%)
Reality: Token still worth ~$10
```

### Vybe's Solution

```
User buys 1000 TOKEN at $0.01 (cost: $10)
Wash trades filtered out
Vybe shows: Actual market value from vetted pools
PnL reflects reality, not manipulation
```

### What We Track

* **Cost basis**: Accurate purchase price from vetted trades
* **Partial sells**: Correctly adjusts remaining cost basis
* **Transfers**: Distinguishes transfers from trades
* **Multiple buys**: Weighted average cost basis
* **Current value**: From vetted market prices only

***

## Deep Dive: Trade History Quality

### What's Included

* Trades from vetted markets
* Full instruction metadata
* Accurate timestamps
* Proper price calculation

### What's Filtered

* Wash trading (same wallet trading with itself)
* Trades in unvetted pools
* Manipulated transactions
* Bot spam (where detectable)

***

## The `vettedMarketsOnly` Parameter

Some endpoints allow toggling vetted market filtering:

```bash
# Recommended: Vetted markets only
GET /trades?mintAddress=TOKEN&vettedMarketsOnly=true

# Research only: All markets
GET /trades?mintAddress=TOKEN&vettedMarketsOnly=false
```

### When to Use Unvetted Data

| Use Case                  | Why                                |
| ------------------------- | ---------------------------------- |
| Manipulation research     | Study wash trading patterns        |
| Market structure analysis | See all pools, not just vetted     |
| New token coverage        | Token may only have small pools    |
| DEX comparison            | Compare activity across all venues |

**⚠️ Warning:** Unvetted data includes fake prices and manipulated volumes. Never use for production trading decisions.

***

## Validation You Can Trust

Every data point in Vybe goes through validation:

1. **Source validation**: Only from supported protocols
2. **Market validation**: Vetted market criteria applied
3. **Consistency checks**: Cross-reference multiple data points
4. **Outlier detection**: Extreme values flagged and filtered
5. **Freshness checks**: Stale data marked or excluded

***

## Next Steps

* [Vetted Markets](./vetted-markets.md) - Deep dive on market vetting criteria
* [Supported DEXs](./supported-dexs.md) - Which protocols we integrate with