Welcome to NexusFi: the best trading community on the planet, with over 150,000 members Sign Up Now for Free
Genuine reviews from real traders, not fake reviews from stealth vendors
Quality education from leading professional traders
We are a friendly, helpful, and positive community
We do not tolerate rude behavior, trolling, or vendors advertising in posts
We are here to help, just let us know what you need
You'll need to register in order to view the content of the threads and start contributing to our community. It's free for basic access, or support us by becoming an Elite Member -- see if you qualify for a discount below.
-- Big Mike, Site Administrator
(If you already have an account, login at the top of the page)
Thanks, vegasfoster. I think I'll begin with the Limit orders first. I'd hate to lose another $1300 or so like I did with the Market orders with Open E Cry, in case the Rithmic data feed with Optimus doesn't make a difference.
Can you help answer these questions from other members on NexusFi?
I made a mistake in using 'InsideBid' and 'InsideAsk' above. Those two keywords deal only with the CURRENT Bid and Ask -- not for Backtesting.
So I changed the program to Buy at the Close +.02 Limit and to SS at the Close -.02 Limit. In the last 10 days (since I changed the contracts due to the Expiration date of the current contract being only 2 days away), the thing has made over $90,000 Net on 2543 R.T.'s. The largest Drawdown has been less than -$2000.
When I forward-tested the old 'Market Order' program for those 3 weeks, it made more each day than when I would close MultiCharts and reload it to see the Backtesting results.
I think this is a good one, if only I can get filled at the Limit prices shown above.
@torroray, the first idea (we have a thread in the Elite forum) was to use these VPS for automated trading, but it seems some of the users are using it for discretionary trading (few guys from Australia, eastern Asia, ...). The amount of data transfered with from a VPS to your screen is smaller than the market data itself , so it makes sense I think.
Some other users has not very reliable Internet connection, and find it safer to trade discretionary from the VPS itself. Some others wants to be able to trade from everywhere, in an airport, in holidays, from day-job (that's bad guys ), or want to collect data 24/7, ...
A quick answer for @torroray: ask yourself,
- where are your orders sent, after a a mouse click, the exchange or the broker ?
- where your market data are coming to your workstation, the exchange or the broker ?
It's your broker (and your broker has a link to the exchange).
easy... traceroute your broker and the VPS location...
find out your broker OE Gateway IP... you will need to ask their support... with that info on hand... just ask @sam028 to do a ping and traceroute to the IP... and also a ping and traceroute to your workstation IP... that will give you an idea of the latency and the networks you will cross... you compare that to your own traceroutes from your workstation to your broker and to the VPS...
and you can make a decision..
it should be noted the main benefit for the VPS will be on the fact that orders resting on the trading platform while hosted on the VPS will execute a lot faster once triggered... you will have to deal with the latency on the "click" from your site to the VPS when entering orders real time, so I suggest you make good use of order management features within the software to ensure the stops/profit target orders are managed by it and not manually.
First a big thanks to Bobbaker for this interesting thread - please keep up informed about (mis)matches between backtesting and real-time trading. There is very little information about this, and I think it will help estimate how good a backtested strategy performs in real-time.
2500 RTs.. quite a busy strategy I must say.
Could you tell some more about your strategy Bob - I don't mean the actual trading strategy but more the implementation of it and how it performs (not in currency, but in speed and latency and slippage). I've read in the thread that you now use 2 ticks limit orders for entry - is this the same slippage you've backtested with or did you backtest with more slippage? Do you experience slippage when exiting, or is CL liquid enough to have little to none slippage. Can you pc and MC keep up with all those trades?
And, as someone who is also very interested in automated trading, what where your three biggest 'errors' (besides the slippages) that you experienced when translating a backtested strategy to live? Perhaps there are some valuable lessons for us beginners in knowing those pitfalls.
I'd love to hear some more about this, (Though I would understand and respect it if you rather not go into specifics)
And your timing is perfect. I tested the strategy Live, Real Time, with Real Money again today. I only let it run for an hour and a half (9:41 Eastern to 11:18 Eastern time). The thing lost over -$1200 Net on 140 R.T.'s.
I had contacted Sam028 to be able to use his computer in New York. So I had very low latency. And I had the best and quickest data feed -- Rithmic. I was really looking forward to this day.
But the slippage got me again. Granted, it was only 1/2 as bad as the first time I tried this on Automatic mode with another datafeed, but it was not good. Most trades were 1 cent to 8 cents worse than what the strategy showed, most of those 4 cents to 5 cents. And that's per side. So each R.T. was clobbered $180 to $200 for just 1 contract.
Yes, I used the 2-cent Range bars for the actual trading, with Limit Orders at 2 cents worse than the Close of the preceeding bar, using the Rithmic data feed. But I kept the OEC data feed for the Signals (but I changed them to 2-cent Range bars also). The backtesting results for this showed +79,803 Net for 1 Contract on nearly 2000 R.T.'s for a 64% Success Ratio in only 4 1/2 days. The biggest Drawdown was only -$892. And these results were based on OEC data all the way (both the 2-cent Trading graph using the near-term expiration month and the 2-cent Signal graph, using the next month's out contract. (Rithmic wouldn't let me backtest at all.)
Using 2 forward-testing days using the Rithmic data feed on the 2-cent near-term expiration contract, it made + 33,543 Net on 480 R.T.'s for an 81% Success Ratio and a biggest Drawdown of only -$241 for the 1st day, and the 2nd day showed + 24,125 Net on 395 R.T.'s. That was good enough for me to try it for Real.
I don't know how my slippage was more than the 2 cents I allowed in my Limit orders. UNLESS the Computer or MultiCharts couldn't keep up. I noticed a lot of signals weren't acted on when they qualified to be acted on. And there is also the possibility that I may have programmed the thing wrong in my Code. It made sense to me and it worked in the Backtesting mode, but maybe in Real Time, Real Life, I missed something critical.
Jura, I am a Beginner myself. If this worked, then I would not be a 'beginner'. I think the biggest 'error' or problem is the fact that I (and most of us) are buying and selling in 'retail' mode. Those who are members of the exchange (or whatever) I think have better access to the prices than we do, no matter what computer or data feed we are using. Sometimes I wouldn't get filled for 5 seconds or more. By then, the price had already had a nice move and I was buying or selling short at the wrong end of that move.
I think Big Mike called it right in one of his posts: "Don't try to scalp Crude Oil." He goes for the Big Runs and that is probably all we can do. Yes, there IS 'Magic' in this game, but we, the little guys, aren't allowed to use it. I really believe that now.
Many thanks to Sam for all his help and to the others who have posted here. I've learned a lot. And I'm not giving up yet. If I can only dream up a program that somehow anticipates the coming moves. Then I would be in (and out) at much more decent prices. Then we'd all be rich. Either that, or I've got to write a program that will trade automatically using somewhat the method Big Mike uses, going for higher highs and lower lows.
I will never give up looking. Someday maybe it will all come together for stress-free, profitable trading.
-- Bob
P.S.: After exiting MultiCharts and re-entering, at 12:40 the Strategy showed a +16,027 Net profit for the day on 273 R.T.'s. And during that 1 1/2 hours that I traded for Real, the Strategy showed +6900 Net profit. And I lost over -$1200 !
), or want to collect data 24/7, ...
There is very little information about this, and I think it will help estimate how good a backtested strategy performs in real-time.