Quote:
Originally Posted by rickroll
thanks utah, will read up on that stuff
it's amazing how many posts from people involved in this stuff over a decade ago all lead to 5d tony
what are your thoughts on all the arbitrage stuff out there now and their influencer marketing
Hi. I do not know what is out there today at all. But I would not trust it. The dude in the video is silly. We had arbs fully automated almost 20 years ago. Stuff has to be way better today. There is no need to sit at a computer.
We built in a full bet sizing and bankroll management system so you could bet full kelly, half kelly, or whatever you wanted. You could arb or simply place the action on the errant line (smarter). We also built in humanization and errant line detection to prevent looking like a computer and to prevent angering the sites, which was my introduction to 5D Tony as he called me and screamed at me for 30 minutes for hitting an obviously errant line. Then he calmed and told me how things work being a pro and how I was expected to act. Then he gave me a lot of free plays for pointing our errant lines to him that I detected.
I also assume the sites have gotten faster and better at updating their lines. Even 20 years ago, most arb opportunities were so fast that a human without software could not detect them. We had a 3 second standard for retreiving a line, placing it in the db, comparing it to find the arb, and triggering the automated betting sequence. And that is not particularly fast. Pinny was always a problem as they used to limit how often you could read their feed (no idea what they do today). So we set up large blocks of IP addresses from places around the world and we would rotate the ips to pull the line.
From 2005 to 2012, I had multiple high level developers working fulltime building and maintaining the software. It is not easy. So I would not trust anyone based on this alone.
Also, you can't really sell or license arb software as it would quickly cannibalize itself. Too many people would be trying to hit the same arbs. Either too many of your subscribers would miss out or you would alert the sites and bring attention you did not want. The way I solved this was by licensing to a select group here. We didn't all hit the same sites or same bets depending on betting strategy. The software detected around 300-500 arbs a day (maybe more. I don't remember, but it was at least that many). So there was enough diversity to never be a problem.
I would post a graph of the results of my first 4,800 automated bets over 51 days when I was proving out the software. The bet size was only about $500 then. The return over the 4,800 was approx $40,000 or 2%. But I don't have posting priveleges to show here. The chart is wild as it clearly highlights the demon known as standard deviation. While I hit our theoretical 2%, I was actually at $50,000 at bet 1900. This means I lost $10,000 over almost 3,000 bets where I had a positive EV. By sheer dumb luck, you can run bad over thousands of wagers. Everyone knows that, but even today it is still weird for me to see.
Probably more than you wanted to know about arbitrage and arbitrage software