Search results

Prop fund might not fit with case-study Harvard MBAs. Might be better off with a number monkey Carnegie Mellon MBA.

Some wise guys who have come to the same conclusions thorughout history. 'Bene vixit, qui bene latuit' [He lives well, who lives well hidden] is originally from Virgil. While these folks were often talking about how to avoid being a target of the Church or some oppressive regime, perhaps we...

I'd like to admit he was wrong on his behalf. Modern portfolio theory makes no assumption that the securities are not correlated. How about discretizing returns for all strategies to the lowest frequency (monthly?) and running your MVO on that?

Elefevre - Any updates on what you all are up to these days? I check into the website periodically but haven't seen much lately. I'd still like to see ETF data - I think ETFs are a favorite of many system developers. I may have some time coming up away from other trading projects to do...

Hopefully, by the time I'm an Oldie, immigration and natural population growth will provide 2-3 Youngies to help pay for me. I think it could sustain quite awhile longer. Not sure how long.

FYI I think there's a way to use tradelink to interface tradestation w/ REDI (and a host of other platforms): http://code.google.com/p/tradelink/wiki/TLTradeStation

Ok, I see what you and TGregg are saying. I assumed it was a 20% back on losses after a whole session or "stay" at the casino. On each hand? that would be crazy.

The 20% doesn't make 10,200 losing hands into winners. It just lessens the magnitude of losing hands. If you plotted a curve of likely outcomes after 100,000 hands, you'd have the center below 0. With the 20% sweetener, the distribution would just shift right a bit. However, the probability...

Bitcoins

Sorry none of this is helping. If you make a new program that doesn't have your strategy and just sends 100 orders (i.e. limit orders away from the market that won't be executed) -- maybe it is something you can share with others running Sterling and get some insight. I switched away from...

I think what you want to do is deconstruct every difference along the path of the order that could conceivably make a difference. What versions of Excel/Office on Computer A and Computer B. What version of .NET on Computer A and Computer B. What OS version installed on Computer A and...

One thing that Algodeal has right is that a lot of this business is about technology and speeding up the process of evaluation. Getting data, backtesting that doesn't have major flaws, comprehensive statistics, etc are all important. They are doing the right thing by focusing on the tech...

Algodeal has a system where they give you a little PDF certificate with an MD5 (or some other hashing function, I can't remember) - which you can use to, ostensibly, prove that you gave them certain code. I'm not exactly sure how that would be worked out to prove when someone stole your code...

Please sign me up at an investment bank with this deal. Looking forward to hearing from you. Thanks. Helloooo? You do realize that I am right here reading this post, don't you?

Cool - are there any asset classes or timeframes you are particularly interested in? Do you generally contact people after a high score or only after you've done some review of their code?

I suspect that you might see more work from the crowd once you start trading strategies and cutting checks. I'd like to spend more time with it, and I probably will, but I'm pretty occupied. I think most folks who can work up a good strategy can work out the technical details of backtesting...

"Bene vixit qui bene latuit" He lives well who lives well hidden. --Ovid

It's only a matter of time before these Arks face stiff competition from DeepBore Networks.

I presume boats would be wired to underwater fiber. Let's assume latency from (A, NYC) to (B, London) is 6ms. You're trading a stock and monitoring PriceA and PriceB. If you're only trading on one exchange or the other, I think you're right about there being no benefit. You'd make your...

Same Sterling servers? (not trying to beat a dead horse, but I think could make a difference) You might put together a test case which sends a bunch of limit orders away from the market and cancels them. I remember reading somewhere that some servers had a limitation of 25 orders per second...