Welcome to NexusFi: the best trading community on the planet, with over 150,000 members Sign Up Now, It is 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)
I keep hearing quotes that 70% or 80% of the volume on the NYSE is computer/ATS trading. So if ATS is so unprofitable, then why is the predominant amount of volume automated? To play devil's advocate, most of this is HFT guys that have computational and latency advantages not available to retail traders, so most of this volume is not using laggy technical indicators or chart/candlestick patterns to place trades like most retail traders do, rather, examining order flow and matching bid/ask disparities across other pools of liquidity coupled with the ability to buy low and sell high somewhere else in sub-millisecond speed. However, there are a fair amount of ATS trading higher time frames out there that work just fine, you're just not likely to find them available for $599 or even $5,999 online.
I know some guys in Newport Beach that sell a forex system for $5,999 a pop with lots of cold calling and they do just fine, though the owner of the company is always "in cognito" and always seems to be on the run, hiding from someone out to get him, so the degree of his "success" can be debated, but he brings in > $2M a year is software sales. Anyways, once sold to, the new client is assigned a broker to help the new client get their account funded and their new software configured with their account and my friend, who is a broker with them, says "That software you bought? It's crap, let me manage your account for you" and bam, he has a new client. And then he trades the account, getting a piece of the spread on every trade he makes. I don't know how well he performs in his trading, he uses some MetatTrader indicators a friend gave to him from forexfactory or some other FX bulletin board (which he sent me to evaluate) and just trades that. If he blows out their account, oh well, he got some nice money from the FCM/IB, but, obviously, he doesn't want to lose the person's money then he won't have any more bankroll to trade with. There are lots of wealthy people looking for places to spread their money around to and blowing up a little FX account with them is not a big deal because they don't blow their whole wad with 1 unknown cold calling broker.
I've also been approached by other people that are IBs or run brokerages that are always looking for good ATS to trade their client's money with. They have systems they use now, and are always looking for more. The ones I've run across mostly have forex systems and a lot of them are looking for good systems that can trade ES because of the available liquidity it provides for trading large position sizes. They say they will take my system and forward test it for 6 months in-house before they will put any client money on it. If they like the results they'll use it, splitting the commissions with the developer. I know a Tradestation trader/developer that has done this with some success. As for me, if I had some great ATS I'd just be happy trading my own account and not pimping it out for a cut of some pips.
The point is, there is a lot of ATS going on out there, and plenty of them are making money.
As far as my bad day, it's 1 day. And I don't follow Jeff's rules exactly so I don't hold Jeff accountable for any of my results.
1) I don't do stop and reverse, don't want to get burned twice by the same bar, but I would like to test how well stopping and reversing actually works. My previous studies of IBs are that there are plenty of IBs that later form outside bars that chop you up on both sides of the bar. Eventually, price will move in your favor away from the IB but you may get chopped up 2, 3, or 4 times before that happens, though with CL during the pit sessions with 15min bars that's less likely to happen.
2) I don't trade reversal bars nor outside bars, just inside bars
3) My targets are a little different and will probably be much larger starting next week, and I'll probably switch to trading 1 lot.
I have made enough "Jeff +8/+16 inside bar" trades (gotta try any system for at least 20 trades) to now step back and look re-evaluate if this is something I want to pursue, abandon, or improve upon. What I've learned so far is
1) While I get 2-3 ticks of slippage, I could improve that with limit orders, but I'm afraid of missing runaway trades that never get filled, and the losing trades I've had would not have been prevented with less slippage so I'd rather suffer a little slippage than miss out on a number of winning trades. And, of course, you get filled 100% of the time on the losing trades
2) I have not seen any correlation in trade performance in the morning vs the afternoon. This should be tested.
3) Trading a system with 2 cars, each with different stop and exit rules is essentially trading 2 separate systems, so I'd rather focus on mastering trading 1 lot with 1 set of rules before adding a set of cars that have different rules.
My interest in inside bars has been around for a while, starting with my study of the DIBS system earlier this year and made some progress developing some indicators for them but never got around code coding a strat for it. I was just excited to see that someone here at nexusfi.com (formerly BMT) was trading IBs in a way that was working for them.
The following 3 users say Thank You to shodson for this post:
Yeah... I am well aware of the HFT trading that goes on and the volume it represents in the market. I know that the Goldmans of the world are out there with automated stuff that works. Retail traders don't have an infinite army of programmers, traders, and money at their disposal. And I don't doubt they are reading order flow instead of an MA.
So when I say an ATS is hard to develop, I mean it's hard to develop with a limited amount of capital and manpower available to the retail trader. I still wonder whether profitable discretionary traders are either breaking their rules to be profitable or have so many rules that they can say they are always following them. Maybe we are arguing some of the same points here.
And hey, I am not saying that the research out there on IB's is bogus and that this isn't a legitimate edge over other methods. I expect to see some losing days. I am just saying I am still skeptical. And I suspect that if you were to employ money management techniques, time of day trading, avoiding the news, trailing stops, etc., to some lame lagging indicator you might achieve similar results.
I don't disagree with you. I've seen a study in Active Trader or Automated Trader (one of those) that had a system that made random buy/sell entries on a market and still made money with sound risk management strategies. I think good entries are all about high probability win rates and good exits are about risk/money mgt, which begets minimum drawdown and good profit factors.
Here are the same tests on CL-08, this time optimizing targets from 8-64, increments of 4. targets 24 and 32 appear to be sweet spots again, with 24 providing the best profit factors and 3rd highest net profit. The 4th parameter is the target size (T1). No break-even stop adjustments. Jeff's +8 did well but not +16, but he goes B/E on those once he gets +8, so going break-even on those may be a good idea afterall. Maybe you should go for +24 on the 2nd car but not go break-even...
BTW, this is trading all day, during lunch and news.
The following 5 users say Thank You to shodson for this post:
Here are the same tests on CL-08 but trading only before lunch, during lunch, and after lunch. Focusing on the 16, 24 and 32 targets:
Before lunch
During lunch
After lunch
Some targets perform better in other times, other targets perform better in separate times of the day. So, while only a few weeks of trading data isn't definitive, at least for this data set, things like "avoid trading in the afternoon" and "don't trade during lunch" aren't statistically supported here but, again, this is a small and inconclusive sample size.
Ironically, lunchtime, at least for smaller targets (8-24) seems to be the most profitable (in terms of profit factor) time to trade IBs. This could be due to a lack of professional traders participating and retail traders, who are more prone to follow a herd if it starts moving in one direction or another, but they don't provide enough volume to push price towards the larger targets.
Maybe the true mix is this: trade medium size in the morning, small during lunch, then trade bigger moves in the afternoon. There.
PS - How can we display an HTML table in our posts? I tried the table button in the advanced editor and followed these instructions but it didn't do anything.
The following 5 users say Thank You to shodson for this post:
I have a friend who does develop HFT systems for a major bank. He is the best software developer I've ever come across.
Whilst he will never tell me what their systems use exactly, he has said that they use 15 or sixteen different "inputs". I take these to mean things like currency and index correlations (and negative correlations). He has also told me that "every system always has a moving average in it somewhere".
The other thing he has told me is that they use large databases for trade decisions. I think they might be storing S/R points, correlations, results of trades and the parameters that were used for each trade etc etc.
One thing is for certain, their systems are not something any inexperienced programmer can knock up in an evening or a trader can develop using strategy wizard.
David
The following 3 users say Thank You to David for this post:
This is a great thread and enjoy reading the posts here quite a bit . My experience with bar pattern strategies is very vast and they are all I use to trade . The reason for this is that they are what I relate to and connect with and for all I know aint the highest prob setup available . Its how each trader uses these setups as an individual that will determine the outcome - they are subjective in that respect . The conditions dictate when to apply these approaches and your instincts need to be sharp to know when to use these approaches and when to pass - more subjectivity . Ive read extensively about IBs and OBs and NRs WRs etc, and have found that what I relate to is always different from what others relate to - the point is that you cant escape the subjective end of this stuff , find what works for yourself .
I commented early in this thread about the similarities between this approach and the DIBS method . Im familiar with this bunch thats trading DIBS and they mirror the results here in that they all have different results in the same markets at the same times but the disciplined traders have the best results overall . Disciplined traders .
all the HFT traders I know don't look at bid and ask, time and sales or any indicators. all they do is enter a trade before a big order hits the market. very low risk trades, of course not available to retailers. so again, the magic "source" is front running and being fast.
I'm sure there are other HFT traders as well, but these are the ones I know.
The following 2 users say Thank You to Silvester17 for this post:
I have tested this with -4 to +4 ticks and I found +1 tick was optimal.
If you look to enter on a pullback at a better price, then you will completely miss the trades that hit their target without ever pulling back. And that has an impact on the results.
The following user says Thank You to cunparis for this post:
