Crypto Arbitrage Platform

Status
Not open for further replies.
alistera

alistera

The fintech I work with are moving away from crypto to pure fund management and are happy to lease or IP transfer the tech, there are currently 20-30 exchange connectors and can run as single timeframe or multi-timeframe, I implemented it for someone a couple of years back which works on core tokens such as Bitcoin/Ethereum but can also arbitrage cross pair such as altcoins.

It's built on institutional tech derived from Forex arb and converted to run with crypto, live arbitrage there are opportunities such as to license out as a monthly service or provide market marking as it out trades Binance, the one I installed was with source code, there is also Crypto Hedge Fund and Crypto Public/Private exchange tech but these are more complex, have regulatory requirements, and are more expensive to both setup and run.

---------------------------------------------​

The platform is an high performance crypto cross exchange arbitrage system using inmemory database to process realtime exchange data at per second intervals and above, auto placing offsetting trades when an arbitrage opportunity arises as the spread widens. The system comes in different options, 1sec, 5sec, 10sec, 30sec, 60sec and above spread basis via 1sec direct from exchange data pulls, the data is then processed and cross exchange spreads are calculated in realtime within milliseconds, when a suitable arbitrage opportunity is found the long/short trades are placed at the offset exchanges automatically, with charting to monitor spreads.

The trades are then monitored in realtime and when the signal is provided the trade is exited, each data pull has a 20% overhead from data collection to trade placement, meaning on 1sec pulls the arbitrage trades are placed within 1-2seconds. Large exchanges such as Gemini and Coinbase are efficient at price arbitrage, however are slow which allows longer spread times in minutes, for increased profits smaller exchanges can be used, these are time efficient however are not price arbitrage efficient where 10sec spreads are optimal.
 
How much profit can it generate ? Because I heard this is very difficult to do this on pure arbitrage as it is already too crowded here.
 
TrAndy2022 said:
How much profit can it generate ? Because I heard this is very difficult to do this on pure arbitrage as it is already too crowded here.
More...

You mean this, it was from when I was installing the platform.

https://cryptonews.com/exclusives/a-rich-man-s-game-crypto-arbitrage-trading-3341.htm
https://99bitcoins.com/bitcoin-arbitrage/

"They [the fintech] are saying you need $100k to $400k to make any real profits in arbitrage with 0.6% return, we managed to get it to 3%-5% return so $50k BTC makes the same profit, but needs more advanced order code due to each exchange having a completely different API structure to stop any trades being left open, so $61k buys the same profits as $100k to $400k".

It is difficult which is why it is based on low latency tech, with altcoins the capital numbers drop by 10x to 100x.
 
Hummingbot is a good system, it's mainly used for market making but you can use it for arbitrage in some cases, CCXT is not great, with both you still need to build the strategy yourself, but yes they are both valid options if you have the months plus of full time development and testing, the ArbXT is different in that it sits underneath the exchanges via performance rather than on top of them.

The simple fact is the system is worth $100,000s built on a platform worth $millions, this uses the same database architecture as top futures and stock exchanges built to standards only the top institutions have, everyone complains they want to have access to something that can complete and produce returns, with the R&D cost already written off as it served another purpose hence they parked the platform, it's an anomaly these consultant don't use it, it's an anomaly that I have access to it, in a bizarre world that everyone has a 'how difficult can it be' approach creating SAAS arb via Hummingbot and CCXT.
 
Last edited:
Websockets fail, it never ceases to amaze how people want to obfuscate something that works, not at this level it cannot (Little League), and anyway it cannot be run as SAAS unless you are willing to pay the $250,000 commercial database license fee (MLB).
 
Status
Not open for further replies.
Back
Top