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)
I am still pretty new to FIO, but love the trade journal idea. Started my membership during the month of the trade journal contest. I was encouraged by @Grantx to start a journal so that is what I am doing now. I had already started this journal a couple weeks ago, in private, as a test of an automated system I am testing/developing. This test is not on a real-money account, that comes later but I am eager to move it over.
Also, I tend to override my system with discretionary exits as the exit-logic is never really perfect. I find it difficult to code every little detail that I can sometimes see just by looking at the price action. Therefore, you will frequently see mentions of whether I would have overridden the system were it on a live account. I like to let the testing account just run, without overriding it, but try to keep tabs on when I think I would override it. The other aspect to this is that I am not constantly monitoring the system, which means that I can only manually override it when I happen to login and see what's happening. On very rare occasion, you just kinda have to know when to use the force. The rest of the time, you have to know to keep your mouth shut and stop thinking you're smarter than everyone.
The basics of the system are: it is a trend following system running on four-hour charts, using ADX as the primary entry and exit signal..though I am not currently going to go into specifics. I use volume as a filter. Takeprofit (TP) and stoploss (SL) are a fixed multiple of ATR. There is also a trailing stop (TS), based off ATR but it does not engage until profit is >= one multiple. The entry occurs when a sustained short-term trend arises on the chart, and the script will exit if that trend peters out. The average duration of a trade is three bars (around 8-12 hours), but the trend really needs to be consistently sustained over six or more bars....which does not happen often.
I was running a prior version of this, but added the ATR-based TS and re-opti'd for several pairs. My parameter-selection process (after I've run exhaustive opti's) is a bit simplistic but I tend to value low-standard deviation profits/losses. The prior script was doing okay but not profitable; it was “just slightly below flat.” My instincts tell me this was merely a temporary blip during a flat time for that system, and that as soon as markets began trending sufficiently (what does that mean exactly?), the system would be overall profitable (remember how trend-based systems work and their basic behavior patterns, duh). Anyway, I re-opti’d several pairs, using a trailing stop with bool input to turn on/off in opti, but with fixed-multiple of ATR for SL/TP. Some pairs have the TS turned off, while others do not. I have set up 13 spot currency pairs to trade under opti’d settings. I feel the addition of the TS and the fixed-R SL/TP is a better plan than using an opti'd multiple of ATR for SL/TP. Certainly, it makes opti runs much faster. Need to figure out a way to automate the opti process. Either way, beginning anew, we will see what these opti’s bring.
One should not expect past absolute-value performance to be an exact indication of future performance. But I think it is fair to compare relative gains to determine which currency pairs ought to outperform other currency pairs. According to my opti runs, the best performers should be in the following order of priority: 1) GBPYJPY, 2) EURJPY, 3) EURUSD, 4) AUDCHF, 5) NZDUSD, 6) EURNZD, 7) GBPUSD, 8) USDJPY, 9) EURAUD, 10) USDCAD, 11) SGDJPY, 12) CADJPY, 13) NZDJPY.
Also important, when I refer to "points" or "pts", I am actually referring to tenths of a pip. So 1 pip = 10 points for purposes of this journal.
And I started running this test on March 12 and have kept my own private. So I will post the old posts here, and then once I catch up, I will post all new stuff here.
My goal for this particular system is to flatten out the equity curve of a trend-based system.
This was posted to my private journal on March 15.
The new and reopti’d system finally made a trade. The trade was a long on AUDCHF, entered at 0.76561 and exited by hitting the TP at 0.76984, for a +423 gain. The price went south upon first entering the trade, on the entry bar the OHLC was 0.76533/0.76645/0.76468/0.76507. The low led to a -65 point unrealized loss and I observed it at around this point in time. I notice it was essentially at a double top formation with a local high from earlier in the week. I also notice that stochastic indicated it was in an overbought area; therefore, I wanted to exit manually but I let it ride. One thing I failed to notice, at that time, was the divergence between the stoch and price action. Although the stoch was high, the bar just before entry (i.e., Bar[1]) had made a higher high on price, but a lower high on stoch. It was obscured because the current bar was still forming, and it was not yet clear that stoch was forming a new lower high. All I saw was a stoch creeping up while price was creeping down. The reality is, today was when the FED announced a new rate hike. The US equity markets made gains, and the USD lost value. Therefore, the increase in the AUD vs CHF here almost certainly was a result of that, particularly since the gain was really made all in one fell swoop of one bar...that is not normal for this system. The unusually large range of the H4 bar indicates that news events (shocks) were at play; plus there was a large swing in the EURAUD in a AUD-positive manner, which indicates a gain for the AUD in light of the FED rate hike. Therefore, this was luck and fortunate timing, but does show the system can pick up on a new trend forming following when the H4 bars go flat for three or four bars. It likely picked up on people exiting the CHF in favor of the AUD, then when the news event hit, more money evidently flowed into the AUD and allowed this pair to hit the TP. I tend to think the TP was set to an appropriate for this pair.
Also, of historical note, both the S&P500 and DJIA have now gone 106 days without a 1% daily drop, tying a 22-year old record; the last time this happened was in 1995. From the news article: “Put another way, Wall Street investors have gone through Halloween, a stunning election victory, weeks of shock, Veterans Day, Thanksgiving, Christmas, Hanukkah, New Years, Martin Luther King Jr. Day, Valentine’s Day, Presidents Day, four ballistic-missile firings by North Korea a tumble in oil prices, with nary a 1% blip” from either index. ...I believe these are unusual times, for sure.
This was posted to my private journal on March 17.
Today was a Friday, and at around just before the last bar of the week, the script entered a trade on EURUSD and USDJPY. Both were shorts, which meant it took opposite sides of the USD trade for each.
The EURUSD short trade occurred at 16:00 chart time, entered at 1.07412 and exited by hitting the SL at 1.07618 (on Monday morning, which was March 20), for a -206 loss. I would have liked to close these positions as soon as I saw them profitable because I do not like opening trade late on Friday and I do not like holding trades over the weekend. Sometimes it can be good, but I think the risk outweighs the potential gain.
The USDJPY short trade occurred at 16:00 chart time, entered at 112.668 and exited by hitting the TS at 112.631 (on Monday morning, March 20), for a +29 gain. This was actually the TS that was trigger because it had gone positive enough to engage the TS. I saw this trade at around +120 pts but did not exit manually because I thought it would continue to go positive.
This was posted to my private journal on March 20.
Today is Monday and the EURUSD trade exited for a loss. The problem is, the script entered a second EURUSD today which also exited for a similar loss. Looking at the screenshot, you can see both EURUSD trades were entered at pretty much exactly the wrong time. These two trades wiped out all the gains made.
The second trade for EURUSD was a long trade that occurred at 08:00 chart time, entered at 1.07699 and exited through the script’s normal exit logic (end of a trend) at 12:00 chart time (only open for 1 bar) at 1.07492, for a -207 loss . . . ironically almost the same loss as the other EURUSD trade from Friday that went over the weekend. As you can tell from the screenshot, the trade went negative as soon as it was entered. This was just one of those trades that should not have been entered. We are clearly in a range bound situation right now and my settings are not appropriate for these conditions. I think I would take this script off if I was running it live, but I also believe that now that the market is solidly in a range-bound regime, the script’s trend filters will keep it out of additional bad trades until it is re-optimized—but truthfully, there is no way to tell for sure.
New idea for a filter: if wick of Bar[1] is bigger than the candle body, then don’t enter a trade...looking at lower wick for shorts and upper wick for longs. I think this would be a bad filter though. Thoughts?
This was posted to my private journal for March 21.
Today, the system entered a short on USDJPY at 16:00 chart time at 111.899 and exited by way of the TS at 00:50 on March 21 at 111.776 for a +123 pt gain. I am beginning to think the TS is an effective tool as it narrows with ATR and only goes into effect once break even.
So this means there was a both a Friday and Monday trade for both EURUSD and USDJPY. Interesting? Correlation? The Friday and Monday trades for USDJPY is shown in the first screenshot.
This was also posted to my private journal for March 22.
Now, look to the second screenshot and check how the price did after the TS exited the trade on March 21. The price kept going in a positive direction and would have eventually hit the TP, though the TP would have pulled in somewhat based on decreased ATR but not sure how much or if it would have been a significant amount. I cannot really complain about a +123 pt gain when I believe the TS prevents large losses down the road. Perhaps after some time, I can take off the TS and allow collection of these larger runs at the expense of larger temporary drawdowns. But for now, I am focused on minimizing drawdowns even at the expense of losing increased net gains. Besides, with minimized drawdowns, I can increase leverage.
This was posted to my private journal for March 23.
Today is Thursday and the system entered a short on EURJPY at 12:00 chart time at 119.774 and exited by way of normal/primary exit logic of the system at 04:00 chart time on Friday March 24 at 119.858 for a -84 pt loss. If you look at the entry bar, you will see the price action initially went in my favor by approximately +450 pts, which would have been a solid gain well within expectations for this system on this trade. However, the initial TP was set over 1,000 pts away from the entry. This is really an unrealistic TP for this, and I am inclined to reevaluate the initial values for the TP, even though the TPs are pulled in as the trade progresses.
As you can tell from the screenshot, I took the photo long after the trade exited, after the weekend.
Late on Friday afternoon, March 24, the system entered a short on GBPJPY at 16:00 chart time at 138.679 and exited by way of normal/primary exit logic of the system at 12:00 chart time on March 27 at 138.745 for a -66 pt loss. In general, I believe entering a trade late on Friday is a bad idea. I am also not a fan of holding positions over the weekend. But I am beginning to believe this is perfectly fine as long as I watch over the trade, and step in to override the system “correctly”...which is perilous.
When I first saw this trade on Friday, my gut reaction was to close for a small loss at around an -80 pts loss (which is just what it happened to be at when I saw). The result would have been, essentially, the same as what ended up happening anyway. However, I let the trade run just to see what happened on the Sunday night open. To my surprise, it gapped down (i.e., profit-positive) and continued running positively for me. I just happened to look at the trade when it was at around +420 pts. The max unrealized profit was about +460 pts, but I was not actually watching the trade when that occurred.
When I saw it at +420 pts, I said to myself I would definitely exit here based on the consolidating range-bound pennant formation on the broader term. The pennant formation can be partially seen with the two red trend lines drawn. In other words, I knew the price was not going to trend significantly lower. The TP was set to 137.722, which was an unrealistic target. Had I exited when I saw the trade at +420 pts, it would have been a tremendous gain over what I actually received, and that gain was something that I could easily see was going to evaporate. Unlike other instances, I do not believe this is hindsight bias due to the consolidating pennant that I saw at the time it was +420 pts. Plus, there were a lot of long wicks on the bottoms of the candles, leading up to where the trade entered.
This was posted to my private journal for trade occurring on March 27.
On Monday morning, on open of third bar of week, the system entered a long on GBPUSD at 04:00 chart time at 1.25283 and exited by hitting the TP at 12:15 (same day) at 1.26068 for a +785 pts gain. This is a situation where the TP was effectively set, not only because we got a high gain, but also as can be seen on the chart, the upward trend ended and reversed soon after the TP was hit—which is basically the ideal scenario. This system does best when we have sustained trends in one direction where the bar sizes are uniform’ish and sustained over a three to five bars. You can see the up trend was sustained over five bull bars, and the entry came at the start of the third bar in the trend. This number is, of course, based largely on the lookback/period setting opti’d for in the system as applied to the particular instrument.
This one win is the largest single win and one of only two that have hit the TP. Moreover, it is one of only two trades (out of a total of eight) that ended with more than +125 pt gain. The old adage of trend following systems remains true with this system—a few big trades make up for all the small losses.
Btw, for the FIO crowd used to NT, the black dotted vertical line represents the beginning of the week and these are 4H bars.
Good luck fivewhy! Ill be following your progress.
I reckon that this is a good filter to use but you would need to give it context. I would say to ignore that condition when the wick is adjacent to other candles and buried in sideways action and the candle body is a certain type (like its a pinbar indicating a reversal to go short but the body has closed positive), but give it high priority when it is protruding out in an obscenely suggestive manner from everything else.
Id also be a bit wary of entering trades late Friday and holding over the weekend. Forex is fine to hold overnight during the week but for me the weekends was a definite no go. Just an opinion though.
