Finally, here's the TS code for the system using 30 min. bars.
Hope I don't lose anyone with this complex system .
Hope I don't lose anyone with this complex system .
You are using an out of date browser. It may not display this or other websites correctly.
You should upgrade or use an alternative browser.
You should upgrade or use an alternative browser.
NG collaboration development
- Thread starter 40yotrader
- Start date
Quote from alpha_monkey:
Hi there,
This looks interesting. Thanks for starting the thread.
I've traded Nat Gas in the past and found slippage to be quite painful. What slippage and brokerage cost assumptions did you make on your tests?
Regards,
AM
40yr,
Alpha has a great point. Your slippage will be substantial when executing realtime. Combine this with bad ticks and NG is a tough commodity to trade without longterm patience and deep pockets.
Regardless of what I say, good luck with the program.
Quote from 40yotrader:
I used $75 per r/t for slippage and comission per-contract. I have no basis for this since I haven't looked at the min. data yet. I chose that amount so I could compare my systems to systems in Futures Truth. What estimate do you think is reasonable?
From the NYMEX website:
Minimum Price Fluctuation
$0.001 (0.1¢) per mmBtu ($10.00 per contract).
However, the contract actually trades in 10 tick intervals. This means the real life spread is $100. As the system would be, in essence, activated at/near breakout points, the liquidity may be thinner than normal. In terms of boiling this down to a reasonable slippage estimate, it's hard to say. Maybe $200, assuming relatively small size (<20 lots)?
Any other thoughts on what constitutes a reasonable slippage estimate for this system?
AM
40 something,
Interesting thread. I've been trading NG for quite a few years and it can be a very profitable market. I trade it end-of-day with a fairly simple system that works a treat. Here's the equity curve (trading a single lot - no compunding or MM):
<a href="http://tinypic.com"><img src="http://i2.tinypic.com/1zvza81.png" border="0" alt="Image and video hosting by TinyPic"></a>
From my experience I can share with you a few points:
1. NG is actually quite a thin market.
2. Like many commodities the front-month has alot of "noise" for obvious reaons. From an end-of-day perspective this is bad, on an intra-day this might actually be a good thing. I'm not sure, I haven't tested it.
3. It can go into into really steep backwardation/contango.
4. Slippage can be a dramatic.
5. It has a large limit so humungous intraday moves are possible.
6. Margin is as volatile as the underlying.
7 This market in not for the feint of heart. Measuring it by HVol NG is the wildest futures market on the face of the planet - bar none!
The other point I would make is that the COT data for NG can be very useful.
If this thread kicks along I'm happy to contribute.
bolter
Interesting thread. I've been trading NG for quite a few years and it can be a very profitable market. I trade it end-of-day with a fairly simple system that works a treat. Here's the equity curve (trading a single lot - no compunding or MM):
<a href="http://tinypic.com"><img src="http://i2.tinypic.com/1zvza81.png" border="0" alt="Image and video hosting by TinyPic"></a>
From my experience I can share with you a few points:
1. NG is actually quite a thin market.
2. Like many commodities the front-month has alot of "noise" for obvious reaons. From an end-of-day perspective this is bad, on an intra-day this might actually be a good thing. I'm not sure, I haven't tested it.
3. It can go into into really steep backwardation/contango.
4. Slippage can be a dramatic.
5. It has a large limit so humungous intraday moves are possible.
6. Margin is as volatile as the underlying.
7 This market in not for the feint of heart. Measuring it by HVol NG is the wildest futures market on the face of the planet - bar none!
The other point I would make is that the COT data for NG can be very useful.
If this thread kicks along I'm happy to contribute.
bolter
Gentlemen, thank you for your valuable contribution. If slippage + comm. ends up in the .02 per-side range then I might have problems. I'm looking at wins in the .10 - .15 range so losing .04 from that would be serious money. I'll do some testing by placing stop orders and try some market orders and see what the slippage is like. If it's more than .02 per r/t then I think I'll have to go for larger moves and fewer trades.
Bolter, thank you for your input. I love the volatility in the NG market. That's why I'm attracted to it. I haven't looked at volumes yet, but hopefully there will be decent daily volumes between Dec. - Feb.. Would a broker providing direct floor access be any help with slippage?
Bolter, thank you for your input. I love the volatility in the NG market. That's why I'm attracted to it. I haven't looked at volumes yet, but hopefully there will be decent daily volumes between Dec. - Feb.. Would a broker providing direct floor access be any help with slippage?
Quote from ImamicPH:
How are we collabing if I have to go buy data?
You need to have a source of data to do much work. As far as I know all data for NYMEX markets is copyrighted and distributed under license agreement. Anyone trading this market should be prepared for swings up to 30k in a day per-contract, so buying historical data for a few hundred dollars shouldn't be a big deal. If you don't trade this market or plan on adding it in the future then this thread isn't for you.
This is just another experiment in working with people on ET. I've tried sharing (as in my first journal when I was a newbie). That was a waste of time. Then I tried swapping systems for markets I already trade. That was also a waste of time. I'm trying this to see if it can be used to improve the speed it takes to develop a working, profitable system. I posted in the classifieds and got many positive responses so I thought many would contribute. Since I've never traded this market I'm open to all input. If this works out, I thought of doing the same for a Euro/USD model.
Who knows, maybe I'll find someone that we work really well together. Next stop...hedge fund (lol, at least I know how to market that). If it turns out to be a waste of time also then at least I tried to work with others. I can always sit around and toil in obscurity (like just about all the other independent traders).
soon to be 44yotrader
Quote from ImamicPH:
How are we collabing if I have to go buy data?
40yotrader's data is copyrighted. This means he can't legally share it.
However, that constriction doesn't necessarily mean the end of a potentially interesting thread. What ways are there to improve on the simple system provided for discussion?
- Use 60/90/120/150... min bars instead of a 30min bar?
- Do double/triple MA crossover systems tend to provide more stable results than high / low breakouts?
- Introduction of an initial ATR stop?
- A breakeven even stop on accumulation of x ATRs profit?
- How long do the best trades last for?
- Use or no use of a profit target?
- An exit rule that deals with excess volatility?
- An exit rule that flattens a position after the market fails to make x ATRs in y minutes?
- An investigation into when it is optimal to add to a winner?
- Multitimeframe analysis - i.e. only take 60min trades when the direction agrees with the daily trend direction?
- Should one run two identical systems that run on different timeframes?
- What other types of system can be profitably deployed?
- Seasonality effects?
- A list of data releases that affects the market?
- An overview of experiences regarding brokers in terms of execution abilities and costs?