Cryptos:  22,485Exchanges:  547Market Cap:  $1,068,388,047,115.6224h Vol:  $59,871,322,686Dominance:  BTC: 41.3% ETH: 18.8%ETH Gas:  39 Gwei

CoinMarketCap: Read what our contributors have to say. This content is provided by the community. DYOR!

Zash Launches 'Wash Trading Detection Service'

NFTs / Announcements

Zash Launches 'Wash Trading Detection Service'

Zash

By Zash

Created 2 months ago, last updated 2 months ago
4 mins read
Zash Launches 'Wash Trading Detection Service'

Introducing Zash's new Wash Trading Detection Analysis across ETH, SOL and BNB.

Whilst wash trading is prohibited in traditional markets, there is limited protection related to NFT wash trading. Be protected.

Yes we know its a problem but lots of questions remain unanswered :

1 - Which specific tokens in a collection should I avoid due to wash trading?

2 - Is there wash trading on other chains? MATIC, SOL, BSC etc

3 - Is there wash trading on OpenSea or not?

4 - Is the wash trading detection algorithm explainable? What happens in the black box?

5 - Is there any specific customisation for precision and recall of the detection algorithm for different use cases?

NFT traders are particularly interested in which specific tokens in a collection suffer from wash trading, rather than the collection itself.

Using our data, we identified that BAYC tokens displaying the highest percentage of wash trading volume.

For token 3221 and 8738, over 60% of transactions were identified as suspected wash trades.

Here’s more data on that. The volume of wash trades versus organic trading volume was higher than 96% for all tokens in question.

With our Wash Trading Detection Tool you can instead identify a BAYC with the least suspected wash trading volume and highest organic trading volume.

A stark contrast to the other end of the spectrum...

Tokens 8585 and 3749 display 0 wash trades 👀

The NFT market on Polygon, Solana and BSC is growing rapidly, and with it, the demand for wash trading detection.

We are the first in the market to analyse wash trading behaviour on BNB chain.

In the spotlight: Mobox

Mobox is the most wash traded NFT collection on BSC with 7% of wash trading volume.

There is very little wash trading activity across other collections, highlighting the saturation of wash trading activity on Ethereum.

🔎 Degenerate Ape Academy

DAA is the collection with the most suspected wash trading activities at 1%.

SOL demonstrates less wash trading than ETH. One reason could be the lack of incentive on these chains as there are no marketplaces like Blur to incentivise frequent trading.

Explainability 👁️👂📖

Users must be able to understand the rationale of detections made by the algorithms.

In contrast to the “black box” algorithm, Zash has developed a  "white box" model algorithm that provides explainable outputs so users can trust and apply our results.

For results with the highest precision, we are employing the 'cycles detection' methodology. We are able to show the users exact paths of the wash trading process.

For BAYC 3221, the wash trading path 0x13d8 → 0x1d36 → 0x7661 → 0x2c2e is explained to the users.

This transparency of algorithm is essential if the users want to accuse some addresses of wash trading activities.

Customisation 🛠️

There is always a trade off between precision and recall. If you increase precision, it will reduce recall and vice versa.

From the recent wash trading analysis article on Dune by @hildobby_, the filter 3 increases recall but reduces precision.

Filter parameters: "If the same address has had 3 or more buys of the same NFT (excluding ERC1155), it will get flagged as a wash trade."

@hildobby_ also chose the parameter 3 here. There is no right or wrong wash trading filter rules and parameters, given that they are appropriate under different use cases.

NFT marketplaces may wish to ban wash traders. Precision is more important than recall here.

Despite evidence of wash trading earlier, particularly centred around Ethereum, Opensea is showing minimal levels of washing trading activity.

Below is a snapshot over the last 30 days, showing OpenSea’s top 5 traded collections by volume.

From a NFT trader perspective, you may try to avoid collections with high wash trading volume.

In this case the goal is to minimise risk and detect as much wash trading volume as possible. Recall outweighs precision.

This is why Zash has developed a wash trading detection algorithm service to provide a token-specific, explainable, customisable, multi-chain supported and real-time API.

It empowers our users to act on timely and correct analysis outputs for their different use cases.

Zash has extensive coverage of the NFT ecosystem with over 50 marketplaces currently indexed and counting...

Hit us up if you want to discuss how accurate data and useful tools can supercharge your product or data teams 🙏

Need NFT data like this? Below you can find our API docs if you want to delve deeper.

Use our API to launch product features faster (like aggregation, feeds or holder pages) or business intelligence / market analysis (think of them like the Nansen + Dune for NFTs).

Our TLDRs

-Wash Trading models are black boxes, our solution is not.

-Ethereum continues to dominate wash trade volume, although other chains need to be checked and monitored. Bad actors will, unfortunately emerge.

-There is a trade-off between precision and recall, the model varies depending on the use cas, customisability is key

- Wash trading on Opensea seems to be OK, for now

As the market continues to expand and regulators become more interested, we would love to see compliance and accounting providers incorporating more wash trading filters in their products and models. Talk to us to launch faster.

3 likes

|
10 shared

Join the thousands already learning crypto!

Join our free newsletter for daily crypto updates!