Starting an investment business

[89826]

Two pieces of advice: leverage works both ways, and if you can't get a good backtest, you are an idiot.

 

[NameBack]

First live trade today with real money! Woo! Went long! Sigh!

Algo seems to be having a hard time getting ahold of what the market is doing right now. The fast cycling from rally to volatility to (what is starting to look like) crash is throwing it for a bit of a loop. Current market environment looks pretty damn close to our biggest dip in backtesting, which had about 17% losses over the course of six weeks.

It's definitely going to be a roller coaster unless the market pulls into full-on crash mode pretty soon, at which point the algorithm should be able to pick up fairly easily on the large unidirectional downward swing. This first position is going to be brutal though, because we'll be holding for one or two more business days after today (assuming no late-day rally today), and I have a feeling we're going to get killed. Down about 0.5% right now.

 

[Sir Humphrey Appleby]



First live trade today with real money! Woo! Went long! Sigh!
Algo seems to be having a hard time getting ahold of what the market is doing right now. The fast cycling from rally to volatility to (what is starting to look like) crash is throwing it for a bit of a loop. Current market environment looks pretty damn close to our biggest dip in backtesting, which had about 17% losses over the course of six weeks.
It's definitely going to be a roller coaster unless the market pulls into full-on crash mode pretty soon, at which point the algorithm should be able to pick up fairly easily on the large unidirectional downward swing. This first position is going to be brutal though, because we'll be holding for one or two more business days after today (assuming no late-day rally today), and I have a feeling we're going to get killed. Down about 0.5% right now.

Did you?

 

[GreenFrog]



Did you?

this.

my portfolio took a beating and a half this week.

 

[Sir Humphrey Appleby]



this.
my portfolio took a beating and a half this week.

I was going to start investing in late August, then the markets had a bad couple of days so I thought I'd ride that out, then they kept going down, then they only had a couple of days of going up before they went down again, had a lucky escape by about two days and that was more laziness than anything.

 

[ter1413]

You're gonna need a BBerg.......

 

[NameBack]



Did you?

Yup! Down about 1.65% so far!

We're thinking we may have started at the worst time possible, but hopefully we'll pull out of this dip soon. It's certainly a good test of our cojones though. I mean, we always knew there would be bad weeks, even bad months now and then -- no algorithm can be good enough to avoid that. So this is definitely going to test our mettle and see if we can keep our **** together while we ride it out.

Or it might mean that the damn thing doesn't work and we're about to lose all our money. That's definitely a possibility too, but I still think that it will perform well.

 

[GQgeek]

One of my best friends made 40k in his first 3 months of online poker. He thought he was awesome and then he lost everything. Some of these posts sound like newbs on 2+2 that think they're a winning players because they're up after 10k hands. Come back after 100k hands and then we'll see.

How long is this test period supposed to run for? What Charley is referencing, to use the poker term, is the variance. Things could be going well for several months and then boom, it all turns to **** all of a sudden. You see it in poker all the time which is why you need such a huge hand sample to be able to say whether you're winning or not.

Good luck though. If you make money I hope you donate it all to Green Peace, Code Pink and the NLRB. They're all counting on you! (and only a capitalist pig would use the money himself)

 

[stevent]



One of my best friends made 40k in his first 3 months of online poker. He thought he was awesome and then he lost everything. Some of these posts sound like newbs on 2+2 that think they're a winning players because they're up after 10k hands. Come back after 100k hands and then we'll see.
How long is this test period supposed to run for? What Charley is referencing, to use the poker term, is the variance. Things could be going well for several months and then boom, it all turns to **** all of a sudden. You see it in poker all the time which is why you need such a huge hand sample to be able to say whether you're winning or not.
Good luck though. If you make money I hope you donate it all to Green Peace, Code Pink and the NLRB. They're all counting on you! (and only a capitalist pig would use the money himself)

Donate yes, Green Peace no.

And honestly you don't know if you can win or not until you try. Just don't leverage up and lose more than you can handle. And don't make bets on stuff that the government can screw you over with (my oil derivatives this summer went so well until the release of oil reserves messed up my sweet / sour spreads)

 

[GQgeek]

^^ Ya, I'm not trying to rain on anyone's parade, but making any sort of assessment of success or failure of a strategy after a couple weeks of trading is pretty much worthless imo. That's really the only point I'm making.

Obviously there are people out there making good returns on this kind of thing but the important thing is can you keep them going for a substantial period of time?

Out of curiosity, how much data was this tested against and for how long (by whatever metric you're using)?

 

[NameBack]



^^ Ya, I'm not trying to rain on anyone's parade, but making any sort of assessment of success or failure of a strategy after a couple weeks of trading is pretty much worthless imo. That's really the only point I'm making.
Obviously there are people out there making good returns on this kind of thing but the important thing is can you keep them going for a substantial period of time?
Out of curiosity, how much data was this tested against and for how long (by whatever metric you're using)?

Oh yeah, you're absolutely right. I can't help but monitor the damn thing, even though I know I should really just sort of sit back and passively manage it with some zen-style serenity until at least the 6mo mark. But, hey, easier said than done! Instead I spend my day flipping between bberg.com and ft.com and watching CNBC and freaking out as if I'm an active trader. It's stupid, but I'm sure I'll get over it after a while. At least the algo keeps me from trading emotionally.

We backtested over the last decade, and we have a lock-in of one year for our funders. Even one year isn't necessarily great long-term proof, and we believe that this model's efficacy will decay with time. However, if it performs in line with our backtested annualized returns over that first year, then it probably means that the concept is strong, and we can build a lot of different models on a lot of different equities and "reboot" them periodically to ensure they remain fresh and effective. That's the long term plan. And we have some funders who are in for small amounts on the proof-of-concept fund who want to scale up quite a bit if things go well over the first 6-12 months.

 

[nootje]

two things:

1: you wont get over it, take it from someone who has been there and deals with several (former) traders.

2: Should have used your own money.

 

[NameBack]



two things:
1: you wont get over it, take it from someone who has been there and deals with several (former) traders.
2: Should have used your own money.

I did use my own money. Or do you mean I should have only used my own money?

 

[nootje]



I did use my own money. Or do you mean I should have only used my own money?

yes

and fwiw, im not trying to be a pessimist or be too negative, but when larger amounts of money get involved it can get very ugly in all kinds of ways. I really do hope youre on to something, but just have my doubts..

 

[Cognacad]

I am quite interested in investing and I am adept at statistics. I invest on my own without statistical aids and it works out fine, but I have always been interested in applying statistics to investments. Unfortunately I don't have time to do so and I don't think I ever will, but the interest is there. I would be interested in hearing more about this project!