New Day trading system for ES

Starting Tuesday, February 20, 2018 we will be trading a new day trading system for the E-mini ES. The system has an excellent hypothetical back test along with a very good walk forward. Of course the “proof in the pudding” will be actual trades posted on C2.

I have purposely kept the fee low along with a coupon for two months that puts the cost of the system at $0. I am a professional trader and algorithm developer in this industry full time since 2004. I have been trading futures since 1968 along with options and stocks and bonds. Now retired but I am still developing algorithms. I have been trading algorithmic systems since the late 80’s when personal computers became affordable.

I will be adding additional algorithms in the upcoming weeks, some long only, some short only, etc. I will be happy to answer any questions or request for data as we go through this together. Just ask

Can you please add the link to the system page?

Tell us about the system, how many Es max, do you scale in. Is there stops on each trades?

I hope this is the link you are looking for. No trades yet, but will start tomorrow

The trading system represented here is a portfolio of two different strategies. Each trades with 1 contract each, with no money management or leverage included. I usually assume $12.5K per ES strategy so that is why the recommended capital is $25K.

The first strategy is one that can go long and short 1 contract only and gets out every day. It does not utilize any stops but over 17 years of data (only 12 years used for training) it has been very responsive to changes in direction. As a most recent example, the system went short on 2/5 at 8:59 AM exchange time and covered at 15:00 exchange time. Obviously, this was very unusual when it made nearly 100 ES points for 1 contract. On 2/7 it was long, but had a loss of $1.6K. I find that for these types of systems, stops cause more harm than good over the long run. One of the reason I recommend $12.5K to $15K per contract is that when I do Monte Carlo analysis (sensitivity analysis) and I am looking at risk of ruin an account of that size has a “risk of ruin” of practically 0.

Now the second strategy is a day trade strategy that is long only. It is for one contract and does have a stop of $1K. This strategy does not trade as often, so the reason for the stop loss. I am going to re-do the development of this strategy without stops but for now will let it run.

Finally, I am in the process of simulating a different DT strategy that goes both long and short with a $1K stop. It will be completed over the next few days to compare with the strategy I have started to trade with.

My long term intent is to have three different strategies trading at the same time in a portfolio. Thus you could have 1 long, and 2 short and the net position of the portfolio will be -1. I will inform the community when I add a third strategy.

Again, all strategies running are for 1 contract only and may or may not have stops. All satisfy stringent walk forward procedures and we are very careful about our statistical analysis.

As for leverage, I am uncomfortable adding leverage to a system beyond some reasonable measure. If the strategies perform as the out period portion shows, I will add that in the future, but for now we are staying with 1 contract only for each strategy, and the maximum position in the portfolio at any given time would be +/- 2. In the next few days, that will expand to +/- 3 if the new development passes the final tests.

link

Was it you or different Jimmy Tann?

with no track record at all, it’s best you just make the system free for about 5 months and let people try it out.

Good catch. Yep it was me. The issue with that system was it was a swing system but only one strategy. That was traded in the real world as one of 9 different strategies in a Portfolio of $6M as a CTA. Without the other 9 strategies, the system remains great but the Draw downs were substantial. Need the others to smooth out the equity curve.

In this new incarnation, I have developed a couple of day trading systems that eliminate the extensive DD concern. Let’s see how it goes.

TT3, you may be right, but we will see. No trades yet, so still a mute point. By the way, two different names come up for me in the forum because I am using two different sign in’s on Collective2 with two different computers. I will watch that in the future, but I’m still the same retired guy

I just noticed that my day trading margin is 25% of the overnight margin, so before any trades are made in the account, I rescaled the starting capital to $10K. As a day trading system the maximum number of contracts that it will trade is 2, so with day trade margins at this level, a starting capital base of $10K is sufficient.

This did something to the chart, but no trades were made, so I think it just lowers the level shown from $25K to $10k. As of yet the starting capital field has not changed as of yet. I felt it better represent what I am trying to do here, and better to do this prior to any trades showing up that will need to be revised.

Thanks for your patience while I get my preverbal “$XXXg” together

Hello,

Your statement is interesting. First, you are using leverage with a $12,500 or $25,000 account trading 1 ES. Second, we just witnessed what happened to those systems that don’t use stops. I’m not trying to convince you to use stops. I’m just pointing out that this is a VERY dangerous way to trade. The book on the ES has been thinner than normal. Even with 1 contract and no stop, investors can lose everything.

I bring this up because it sounds a little misleading that you’re not using leverage. Investors should understand the risk and trading. Even if it’s only 1 ES contract on an account of $12.5k it is absolutely leveraged up. With risk management in place that’s one thing. You stated you will not have any money management.

If you made the statement below, you may not fully understand what you’re trading. 1 ES had a value today of over $135,000. So you are leveraged. The value of ES is 50 times the price of ES. It seems you think ES is not leveraged at $2700. That’s completely false.

Good day

PayoffMatrix:

You are correct that I miss-spoke about leverage. I do know what the ES contract is worth, since I have been trading it since 2002, but I was using the full overnight margin requirement when I calculated the leverage. When I traded professionally as a CTA, we assumed $25K per ES contract, which at the time was 5X the exchange minimum overnight margin.

As you well know, futures contracts in reality have no intrinsic value themselves, only the cash contrac does. The margin that is offered is actually just a “good faith” deposit that you will pay the same day any losses come daily settlement time. At the settlement time, all holders of the contract must be balanced with the clearing exchange.

Again, I am sorry for any confusion. I was directing my comments to exchange minimum margins. Now just to say that I’m really crazy, after I wrote that post above, I decided that I would decrease the minimum account trade size to 10K for this day trading portfolio. Most brokers require 25% to 50% margin of the exchange minimum for day trading, so I reduced the recommended account size to $10K.

So versus my day trading margin (good faith money) the minimum account would be 3X the broker day trading requirements. Probably a mistake but I can’t go back an increase it back to $25K. The returns both negative and positive will be amplified in this account, so again, probably a mistake.

As to stops, I was just providing full disclosure to the community. I did trade this program through the early part of February in my own account and did have one day that was a $1.7K loss, but in the same week had a $2.5K gain. That was the first time I have ever made 100 points on one day trade contract in one day.

Further, I do not agree with you on stops. Every time I use them in the past I have been screwed. That is why I recommend $25K per ES contract as trading capital if you are going to stay overnight. Here is just one example. I typically develop trading systems over a 17 year data span, with on 60% of that data as the training portion. The other 40% is utilized as follows: 25% for out of optimization testing and then a final 15% of the data that is never used in development but used as a validation segment on the final review of the strategy.

I typically throw away any strategies that don’t have similar equity curves in the testing phase (25%), and meet typically stringent Monte Carlo statistical testing sensitivity analysis and many walk forward tests.

The final test is to use the validation period to see how it performs over that. In this environment, a typically day trading system over the period without stops has a performance of around $165K, a PF of > 2 and a % profitable of > 60% with a maximum DD of 3.3K over its life. Obviously with a 10K or 25K account size the DD is well within particulars. If you want a DD of 10% or less, then allow 30K capital for this strategy. I am not sure about this strategy as of yet, but the proof in the pudding is to actually trade it with real $. That is what I am doing now and have included Collective2 community to follow along.

The best that I have achieved in the past with a day trading system that utilized a stop of $1K, has been half that total performance, with similar PF and % profitable and a similar maximum draw down, so I see no benefit to using stops.

I will continue to trade this in real time to measure its actual real results with the historical data and if it fails, it fails.

Thanks for your comments and hope this gives you some indication of my thinking behind trading E mini’s.

Ups!!! So much for all of the statistics that I used to develop this system. At least we got the worst trade out of the way as a first trade!!! Even though I never develop systems with any trading days excluded, I usually stand aside on a FOMC reporting day, because of exactly what happened today. Oh well, at least you can see all of the bad with hopefully the good coming. Ugh!!! A $1K stop would have saved me today. Well, you can’t win them all and we are in this for the long term, I hope.

One thing, the market can always outlast your money, so watch and learn. I may just go down in flames, but even someone with many years of trading can make mistakes. Don’t trade on FOMC reporting days. It is a 50-50 proposition regardless of any statistical work you do.

@JimmyTann2, if you based your statistics on simulated results then you can pretty much throw them out the window since the markets will change and you will find out that all trading systems will have losing periods and winning periods. The better systems are those that can survive the losing periods without much drawdown.

Also, stops are extremely needed unless you can sit there and watch the trade since there will always be days where the market will drive strongly when you least expect it.

Anyways, they always say that the markets are the best teachers…lol

Hello @JimmyTann2

Thank you for your explanation. I just didn’t want any investors being mislead. Reading that no leverage will be used would lead some to believe that literally no leverage is used. If they deposit $10,000, they can literally lose everything even on 1 ES contract if there is no risk management. Investors should be aware of that risk. I’m not debating your philosophy on no risk management even though I don’t agree with it. This was just a post to help potential investors understand there is risk especially when there’s no risk management.

The majority of investors here use Interactive Brokers. The current day margin is $4271.75 and maintenance margin is $3417.40. Recently IB had their day margin around $8000.

The CME has the margin set at $5800. Brokers have a right to ask for more to protect their business and IB has done that.

I wish you luck on your new endeavor. I wouldn’t have responded to this if you were simply promoting your system. What triggered the response was the “no leverage” statement. No hard feelings. I just know how much some investors lost recently and wanted to highlight and clarify the risk.

Good day

1 Like

AlgoSystems:

Of course the future is never known as we design algorithms. I agree with you that all systems have DD periods and you have to survive them (consequently the “risk of ruin” statistical analysis) but that is what we have to work with. There are numerous ways that one can minimize the future surprises and that is two fold. The first is make sure you always have a section of your historical data that you never never use for any development purposes. I usually set aside 15% of the data for this purpose. It doesn’t really matter where that data is, so you can reserve it at the beginning of the series, the middle, or any where you want. The goal here is to have any algorithm at least perform to some minimal standard in the “out of testing” data. Some developers use a 50% figure of merit for this, I use a more rigorous test and want the equity curve to be similar in nature as the in testing period.

I do divide my data into three parts, but the first two are used for development and then a test period to check each different algorithm or parameter space against. With this approach, I throw away a ton of systems that others may find useful.

The second way you try to look for the unknown future is that you use what is called Monte Carlo techniques usually at a 95% confidence interval and actually, change the order of your data in a random fashion or change the various parameters of the system (sometimes called sensitivity analysis) and also change the start dates of the data test. I usually run 5000 iterations, but in no way do I think this gives me a picture of the future. I’m just trying to exercise the algorithm in any way I can think to make sure it doesn’t immediately fall apart in the future.

As to your comments on stops. All I can tell you is that is a personal decision. I have actually been trading futures since 1968 and learned from my Pop about stops since he traded in the early 30’s up to the late 60’s. If you notice that is experience through the depression, through the boom 60’s, the Hunt corner in the silver market, the cuban missile crisis ( you should have seen what that did to sugar) countless wars, 1987, the tech boom, 9-11, 2008-09 the real estate crash and now the Trump market miracle. I pretty much know first hand how many “black swan” events can take place. There are two ways I have successfully migrated those waters. The first is I make sure that I only “play” with 5% to 10% of my total investment capital. I used Futures as an allocation in my portfolio, just like any other allocation. The second is that I treat margin as what it is, just a good faith measure that clearing exchanges and brokers demand. So, I never trade with the broker minimums (I see IB raised their minimums dramatically on the E mini’s lately but TradeStation did not) . I try to use around $25K per ES contract.

What that does is negate the need for a stop loss. When I do my statistical analysis, there is a 5% chance that my “risk of ruin” will be greater than 5% but that is calculated on a much lower margin base than I actually use. Thus, yes there is a statistical basis that I can wipe out, but I still have trading capital behind it so in reality I’m not wiped out.

If you only have $10K as your total capital, you can’t trade futures with a strict stop. I just manage to do the “stop” function off line with a much more generous capital allocation. On a $100K capital account I would never trade more than 4 ES contracts, and usually only trade 3.

As I write this comment, It seems that my two day trading systems really are not doing well. The second trade is as bad as the first, and is closely meeting my “get the hell out of dodge” level. Probably will do something after the close today. Ugh.

OK, here is the status now. For my second trade on my day trading system is almost as bad as the first signal. Its getting close to my emergency “get the hell out of dodge” when the actual trades are meeting the historical DD even with all of the statistical analysis I employ. So use this as a lesson. Even someone who has been trading for over 40 years can get wiped out with a “sure fire algorithm” that looks almost too good to be true.

I was trying this day trading experiment since I don’t develop day trading algorithms usually. I am a swing trader, but thought this would be an interesting challenge. Well it sure was. Again, it re-confirms my history that I am one of the worst day trading traders in the world and I better get my butt back to swing trading. Something I understand well. I will be sending out a blast to followers that are just simulating this (Phew) and I’m the only one losing real dollars. Live and learn, but if you want to buy at the high and sell at the low, follow this DT system closely. Ugh.

Robustness of this back test in question IMO. Blown up in 2 days?

C’mon bro. Ain’t nobody got time for dat!

@JimmyTann2, i don’t want to lecture you but my advice is (as a former daytrader myself) to move your stops to breakeven after you have at least 5 pts profit. It works! Also, its good to have a target. Your trade yesterday had at least 10 pts profit from your original entry so I’m not sure what you were holding out for since there are swings in both directions during the day session and we rarely get a one directional day.

Anyways, thats my suggestion…you can take it or leave it…lol…

What I really don’t get is why not to trade without any announcements for couple months (at least) and then reveal the results (if it will be results)? Why do you need to throw yourself into this excuses?

1 Like