Seeking other e-mini day-traders

[MBender]

I'd like to get a dialog going with other e-mini traders. You might notice from my earlier postings that I'm working on the e-mini Nasdaq and Euro.

Anyone interested in sharing ideas, tricks, etc. please let me know.

[Doji3333]

I'm here if you want to discuss things. I trade the ES, NQ and EUR also (as well as other stocks and options).

You can check out my new adaptive strategy in the "Trading Strategies" section of this forum. I'm trying to test it and make sure it works properly before turning it on LIVE.

[MBender]

Thanks for getting back to me

Hi Doji3333:
Thanks for getting back to me. I was beginning to think I was the only e-mini trader in e-signal.

I, too have come to the conclusion that the e-mini's need an adaptive strategy. I've been hooked on the Advanced GET techniques, but they tend to work in trending markets and end up being traps in flat or highly-volatile markets. Day-of-week, time-of-day and financial news seems to affect these things (the e-minis) so much that it's difficult to make money. I'm assuming it's that they're not heavily traded, therefore less fluid than non-mini-futures or other instruments.

I wrote algorithms that trade on the Advanced-GET strategies, and I have a market analysis routine that stops the trade under unfavorable market conditions. I'm still toying with the algorithms, so I'm interested in your findings.

I admit I'm not quite sure what you're using for your entry and exit basis, but the charts look good.

Also, how are you getting trade confirmations? or are you?

Cheers

[arbifox]

Trading the E-Mini

I have been trading the Russell E-Mini going on about 3 years through a multi-time frame strategy. 2010 Q1 was stellar, but Q2 erased the profits of Q1, and Q3 has been flat to date. Last year the system closed out at about 145% gain.

I can speak on the subject for days, but the one thing I know is that any system has losing periods but needs to be evaluated as a whole over the course of an entire year.

I have talked to other traders that have a bad month, and then they think it is their program which they quickly change, and they are forever chasing the rabbit hoping to find one program that can make 50 points in a day.

I am currently upgrading my code to version 8 (my system just keeps evolving and I have a custom built backtester). I have contact with Brad from time to time, and I would love to turn what I have into a business, but the results this year really do not lend to me moving forward just yet. Maybe after the results of 2011.

I am always up for conversation and strategy discussion, and the best advice I can offer is to be patient at this game. It takes methodical testing and an understanding of how to design a system that plays the probability of a trade working out, and not always the one off situation when one trade captures a rocket move in a particular direction.

All the best,

Thomas

[MBender]

Hi arbifox

Thanks for getting back to me.

I sincerely appreciate your perspective. I've been playing with this off and on for about a year; but only went serious (full time) a few weeks ago. I liken it to a casino. A casino gambles just like their clients, but they make millions on the 2% advantage. I'm perfectly fine with that.

I'm okay with a bad day, or even a bad week. A bad quarter would be difficult -- I don't have deep pockets. I'm only trading 1-2 contracts at a time until I get more experienced.

You mentioned multiple time frames. Do you find one better of the others during certain market conditions?

I like 1-minute on the NQ and ES. They get you into trends early and at a low risk. When they start to flatten, 2-minutes works well. Brad mentioned 15-minute intervals at one point. I find that time-of-day has too much affect when you go that long. What time frames do you like?

Thanks again.
Michael.

[Doji3333]

Re: Trading the E-Mini

Thomas, great to see you here. It's been a while since we chatted.

I can vouch for Thomas as a serious developer and efs expert. He is as skilled, if not more, than I am in regards to developing automated systems. But I like to think I'm pretty good at it too. :P

With regards to order filling, I use NinjaTrader or my own utility to fire orders to the brokers and I simply need to wait to see if that order fills, partial fills or NO FILL. It's actually pretty simple to do this in EFS with DLL calls.

There are tricks as to how you want to enter the trade and how long you are willing to "work the entry order". I've developed some tools that allow the order to be worked for a period of time and other mechanisms that allow the order to be worked while the conditions are right. If the order fails to fill, then the trade is missed (unless I get another trigger on the next bar).

My adaptive strategy is something I've been playing with for a while. The concept was to build something that could adapt to the market and stay out of flat days (like today). So far, my adaptive system has not taken a trade today - which is what I would expect.

My other system has done really well this week. I run it in a very contained mode which limits risk. It shoots for between 1~2 pt targets on the ES and NQ and is limited to two trades per day. On Monday, two quick targets. On Tuesday, two quick targets. Today, one quick target and currently long (with a target at 1144.50). So I need about 1pt more to the upside and the system will be at 100% accuracy this week (running at 2 pt targets).

I can't tell you how many people tell me "all I need is 2~3 pts a day on the ES with a 70% accuracy and I'm fine with that". Getting to these levels with consistency is not as easy as some people think.

Additionally, as Thomas stated, having a DrawDown plan is key to long term success. I like to use the 5, 10, 15% rule. If my system enters a drawdown phase and reduces my total trading account by more than 5%, then I'll dramatically reduce the number of contracts I'm trading (or stop trading). If it continues to draw down to the 10% level, then I'll again dramatically reduce my exposure, or stop trading. The 15% level is my MAX level to STOP TRADING.

The reason I stop trading at any of these levels is because protecting the equity in my account if critically important, more so than trying to recover from the losses. I expect I'll live a few more years and with that I'll have lots of opportunity to trade again tomorrow or next month. Getting greedy is not the objective, the objective is protecting capital and providing the best opportunity for profits.

When would I turn my system back on, after 1~2 weeks of continued profits with an accuracy ratio greater than 60%. This is one of the tricks of running an automated trading system. I've seen many many systems do really well, then enter a drawdown period where the user just "lets it run". The end result is destruction of trading capital. I've taken the necessary steps protect trading capital and I advise others do the same.

Originally posted by arbifox
I have been trading the Russell E-Mini going on about 3 years through a multi-time frame strategy. 2010 Q1 was stellar, but Q2 erased the profits of Q1, and Q3 has been flat to date. Last year the system closed out at about 145% gain.

I can speak on the subject for days, but the one thing I know is that any system has losing periods but needs to be evaluated as a whole over the course of an entire year.

I have talked to other traders that have a bad month, and then they think it is their program which they quickly change, and they are forever chasing the rabbit hoping to find one program that can make 50 points in a day.

I am currently upgrading my code to version 8 (my system just keeps evolving and I have a custom built backtester). I have contact with Brad from time to time, and I would love to turn what I have into a business, but the results this year really do not lend to me moving forward just yet. Maybe after the results of 2011.

I am always up for conversation and strategy discussion, and the best advice I can offer is to be patient at this game. It takes methodical testing and an understanding of how to design a system that plays the probability of a trade working out, and not always the one off situation when one trade captures a rocket move in a particular direction.

All the best,

Thomas

[Doji3333]

Re: Hi arbifox

Originally posted by MBender
I like 1-minute on the NQ and ES. They get you into trends early and at a low risk. When they start to flatten, 2-minutes works well. Brad mentioned 15-minute intervals at one point. I find that time-of-day has too much affect when you go that long. What time frames do you like?

Thanks again.
Michael.

My main robot, AWTS, runs on 5 minute charts. The only reason I'm running my adaptive system on 15 minute charts is because it is designed to avoid congestion. On a 1~5 minute chart, you can have many many bars/hours of congestion when the markets are flat. Because of this, I chose to use the 15 minute charts because they graphically reduce the congestion on the charts and provide very good trending signals when the markets really trend.

I've used a "rabbit hole" example in past posts, but remember, if you're hunting rabbits, you should not be looking in the trees for them - you need to look on the ground near "rabbit holes".

BTW, I love chatting about this stuff and helping other people develop their projects. More later.

[arbifox]

I use a 3 minute, 450T, and 150T time frames. I also vary the control of the program between the time frames depending on the strength of the trend. This seems to work well over the long haul, but there are times that the market is so choppy, that you just get shaken out.

In general, I assume that the market is going to be choppy and I tried to ride the waves of the day. Using longer time frames can get you into tons of issues if the market is not really trending for that day.

In general, it is really playing the odds. My trading improved drastically when I approached it in a manner that I am looking for the highest probability of a win, and not to capture each and every move. With that in mind, one can always analyze their trades given the time of day and measure the percentage of wins vs. losses. That actually helped me weed out some low probability trades.

Regards,

Thomas

[MBender]

Thank you both

You've both given me a lot to thing about.

I have three basic routines. One is a simple pivot detector. A second is a market analyzer, which determines if the market is trending, channeling, or topping/bottoming. The third is the trade strategies. I have three strategies at the moment, 2 are standard Advanced GET -- the XTL (very easy to code) and False Bar Stochastic (more difficult to program an I expected). The third is my own and relatively easy.

My overall approach is to use the market analyzer to adjust the probability of each trade, and therefore, the number of contracts at risk. I'm thinking of adding time-of-day to the mix.

The three routines work pretty well, I just haven't integrated them yet.

Thanks again.

[arbifox]

Couple of ideas . . .

Thanks for the nice words Brad, I have just been busy with work these days, so less time for development.

The other item I could think of that might be really helpful is to record the state of the main indicators that determine trades at the start of a trade, and see if that can be used as filter for low probability trades. I have encorporated this concept into my backtester, and it does help weed out some trades which add nothing to the bottom line.

The good part to all of my work is that I have a backtester in place that lets me develop without risking any capital. Having a true backtester is key to really getting ultra profitable, and before I had it, I just made arbitrary changes without feeling the effect over the long haul.

It all just takes so much time . . . that is the other caveat to all of this.