SanMiguel, I agree the stops are very large and I would love to try your back testing file but when I try to open it it says access to the file is denied. Do you know how I can access the file?
So I ran through a quick back test of 5 pairs, EUR/USD, AUD/CAD, GBP/CHF, USD/JPY and EUR/NZD.
I did it because the last 4 pairs represent all the currencies I trade and EUR/USD is the first thing I think of when I think currency.
I added one additional entry rule, the Kumo furthest to the right, (what you can see) must agree with whatever trend you are entering or exiting. To exit you must also wait for a breakout from the Kumo.
Here are the stats on the trades.
From 2/27/09 through today:
Pairs: 5
Trades 14
Record: 8-6 (57%)
Avg stop: 479 pips
Avg loss: -290 pips
Avg gain: 419 pips
Profitable pairs: 5
Losing pairs: 0
of stop outs: 1
By profitable or losing pairs I mean you compound the wins/loss % to get the total % capital gain/loss for the pair.
Here is how the pairs came out.
AUD/CAD: 16.6%
EUR/NZD: 6.8%
GBP/CHF: 7.1%
USD/JPY: 5%
EUR/USD: 2.6%
[B]Total gain [/B]from 5 pairs, (not counting compounding) [B]38.1%[/B]
If you compounded the gains from each pair, (meaning one trade closes, you realize gains, then roll them into next trade) you would certainly be above 40%, but lets be conservative and say 40% gain in 8.5 months.
That comes to an [B]annualized return of 57.8%[/B] ON JUST 5 PAIRS!
This system seems to be fantastic for position traders who literally want to check the charts once/day and just set it and forget it.
By using T/K cross, Kumo breakout, Chikou span confirmation AND kumo cloud trend, (bullish or bearish as represented by color of Kumo) and setting big enough stops, you can trade with a trend and only exit when that trend truly ends.
I think there were only 2 fakeouts during my backtesting, so 6/7 trend reversals predictions by the Ichi were correct!
Now with 3% capital risk, and large stops, (ie lower leverage) on each trade, one uses 6.8% of margin/trade, (including maximum loss). So with this system one could have 14-15 simultaneous trades open!
So with 21 pairs that one can trade on Oanda, (23 but 2 are ranging) 14-15 simultaneous trades is very doable, (since each trade lasts on average 3-6 months).
So if the performance this year on the five pairs I backtested is any indication, with 14 pairs, you would have 43 trades/year and achieve 160% annual gain, and again, that is not including compounding with which the returns can go as high as 200% annualy.
But lets be conservative and say that, including taxes, gains of 100% annualy are possible with this system.
Now regarding the Ichimoku system that I am following with this log, it actually turns out that all the trades I made this last week were in accordance with the above described system!
So I donât have to start over and I have more confidence than ever that, despite only 3/9 current trades being profitable, that things will turn around next week.
During my backtesting some trades that became profitable were down as much as 5-600 pips at one time (EUR/NZD short had a stop of 988, went as low as -700 pips but ended up as a 3470 pip gain, total return on capital 10% in 6 months)
Added 4 more pairs, USD/CHF, GBP/USD, AUD/JPY and USD/CAD.
Overall these 4 pairs had a lot more fake outs, resulting in far more trades, (27 trades) Of these only 6 were positive. However, these positive trades where so large that in the end that 2/4 of the additional pairs were positive.
Hereâs the stats for the pairs: (how much your account gains or loses since about march 1 if you traded just this pair).
AUD/JPY: -1.85%
USD/CHF: 7%
USD/CAD: -6.8%
GBP/USD: 7.43% (only 1 out of 6 trades positive, but that tradeâs profit was 18.9% of account value)
Overall the stats for the 9 pairs:
Trades: 39
Win/loss record: 16-23 (41%)
Avg stop: 374 pips
Avg profit/trade: 154 pips
[B]Cumalative performance[/B], (adding all the pair compounded performance) [B]45.88% capital gain[/B].
So its still a winning system, but apparently not as fool proof as the first five pairs predicted.
I will test 3 more pairs tomorrow, 2 pairs on monday and then 1 pair/day until I get through all 21 and of course, I will report the results here.
It may also be true that certain pairs are just better suited to this system?
Has anyone whoâs used the Ichimoku Kinko Hyo for trend trading noticed that certain pairs seem to work well while others seem to offer nothing but fake outs and losing trades?
I have been warned by San Miguel that EUR/CHF is a bad pair, something my own backtesting confirms, so maybe there are ideal pairs to focus on.
Thoughts?
Try again with this:
Just be aware that when testing Daily timeframes, you will probably get some insane results from 2008 onwards and APril 2009 onwards. There have been some strong trends on most pairs (eg AUDUSD, etc.)
TradeSim.zip (491 KB)
I really am not sure how you are getting that massive stops, but I suppose it comes from trades where you get a TS/KS cross that is far away from the kumo?
If I was trading kumo breakouts, any trades after that point would be an âadd on tradeâ for me and I would place my stop behind the Ks - worth looking at as this will give smaller stops. ie you want to get in with your first trade at the kumo breakout, then any trades you add on after that are for extra profits along the way with tighter stops. You should also be moving your stop to breakeven after a certain amount of pips in profit.
Thanks for the advice SanMiguel, yes the big stops where after kumo breakouts. I have noticed that many of the succesful trades didnât require big stops, so perhaps smaller stops could work. Of course smaller stops also means more realized losses, which means risking a smaller amount/trade, thus making smaller gains for the big trends.
Its a delicate balancing act. On one hand the big stops stops give you the certainty that you wonât get stopped out due to daily variation, on the other hand they limit your gains.
Of course, placing smaller stops that get hit often means 1% risk vs 3% which means each win is 3 times smaller and there will be more losses. I wish there was a way to have a computer backtest systems for me but I am not a tech savy person.
Perhaps, a strategy I could use is to wait for a Kumo break and then act, thus allowing me to have smaller stops. (Recently USD/CHF went bullish with just 128 pip stop and EUR/USD short had stop of 180 pips), I believe AUD/CAD short also had small stop of about 100 pips.
I will be sure to note the performance of the top 6 trades in the system I am tracking, which where older trades vs any new trades, (which are fresh Kumo breakouts).
Or better yet, I may try tracking one other system, where you risk 1% of capital, place your stops at Kinjun-Sen and once your profit is equal to your stop, place a trailing stop at break even.
This will result in more trades, more stops, but also more profits, (I noticed that often during a fakeout, the price does move in the indicated direction initially) so maybe my win % could increase to say 60%, which would make up for the increased realized losses, (which are only 1%) and the smaller gains but more frequent gains.
12/15/09 saw a sell signal on EUR/AUD. This brings the system up to 10 simultaneous trades, still plenty of margin left, (40% of account including maximum losses on SLs).
Today was a good day, with the rally in the USD bringing large gains in several pairs.
Still not making up for the 3% SL on EUR/NZD but all my pairs are now moving as they should, so I am very optimistic about the future.
Stats:
Open trades: 10
Closed trades: 1
Win/loss record on closed trades: 0-1
Win/loss record on open trades: 5-5
Realized gain: -3%
Unrealized gain: 2.1%
3 Shorts today:
AUD/USD
CHF/JPY
EUR/JPY
1 closed trade
Closed trade record: 0-1
Realized profit: -3%
13 active trades:
Active trade record: 8-5
Unrealized profit: 8.1%
Things are really starting to move as the Ichimoku predicted.
nevermindâŚ
AUD/NZD long position was closed today at a 310 pip loss, 1.89% loss of capital, because the Ichi indicated that the trend is now bearish. Waiting for Kumo breakout to enter a short position.
USD/JPY long position entered as all 4 criteria for a long position are now met.
Overall Stats:
Closed Trades: 2
Record: 0-2
Realized Profit: -5%
Open Trades: 13
Record: 8-5
Unrealized Profit: 16.6%
Final 2 trades today, CAD/JPY long and GBP/CAD short.
This brings the total number of open trades to 15, maxing out available margin while guaranteeing that no margin call occurs,(not quite, if every trade does to SL, Iâll be $12 short of the minimum margin requirement, on a $1000 account). However, it looks like NZD/JPY short will soon be closed out for about 1.2% loss, which will make my account safe again and allow me to replace that pair with another, again bringing the number of open trades to 15.
So far here are the stats:
Closed trades: 2
Win loss record: 0-2
Realized profit: -5%
Open trades: 15
Win loss record: 11-4
Unrealized profit: 17.7%
Then here are the stats for San Miguelâs system, (1% risk, set SL at Kijun-sen and install TS when profit surpasses initial stop loss size, (buy price + 10 pip profit)
Closed trades: 2
Win loss record: 0-2
Realized profit: -2%
Open Trades: 12
Win Loss Record: 9-3
Unrealized profit: 5.1%
Number of TS in place: 4
Guaranteed pips from TS: 896
Again as a reminder for the rational of the San Miguel system:
The smaller risk is because I am using the Kijun-sen to place my SL, these stops are smaller than the other system where stops are placed 20 pips above and below the kumo cloud.
Because of the smaller stops trades are more likely to result in realized losses, so capital protection is paramount.
Also, since TS are likely to decrease profits on succesful trades, the % of capital risked/trade must be lower.
Note, the San Miguel system has only been running since Dec 13 vs the original Ichimoku system which has been running since Dec 8.
Here is a brief comparison between these 2 systems and a third and fourth hypothetical San Miguel system where the capital risk is 2 and 3% respectivly%.
Original Ichimoku: Realized gain+ unrealized gain: 12.2% over 13 days=.88%/day return (including weekends)
San Miguel 1% risk: Realized gain+unrealized gain: 3.1% over 8 days= .38%/day
San Miguel 2% risk: Realized gain+unrealized gain: 6.2% over 8 days=.75%/day
San Miguel 3% risk: Realized gain+unrealized gain: 9.3% over 8 days:= 1.2%/day
Now the comparison between hypothetical San Miguel systems can only go so far, since the maximum number of trades/system will differ. The 1% system has unlimited open trades, 2% is close to unlimited, since it can handle 23 trades without a risk of margin call, (only 23 tradable pairs) while the 3% system can only handle 15 trades max.
One thing is certain and that is that the Ichimoku Kinko Hyo has so far been very effective at prediciting trends. Now the only question is what is the best way of maximizing profits using this powerful tool?
How best to set a stop loss?
To set a TS or not? and if so when and how big?
What % of capital to risk/trade?
The trade offs between the two systems mentioned above are as follows.
Original: Larger stops= less stops due to news, daily volatility BUT smaller % of winning trades, more draw down risk
San Miguel: smaller stops= larger leverage/trade, larger % profit/trade, BUT smaller return due to less leverage
San Miguel 2,3% systems: higher capital/risk = higher risk of drawdown but also larger % returns if this system can return more profitable trades than losing trade BUT larger risk of draw down especially since stops are smaller.
I only have time to track the 2 actual systems, but until I get to 15 on the San Miguel I can compare the other 2 hypothetical ones, and the 2% San Miguel can be compared with the 1% San Miguel from here on out.
I will assume that the long term returns of the higher 2% compared to 1% is indicative of trade off between leverage and drawdown and thus 2% can act as a proxy for 3%.
Now if only I could determine what system to use for my $10 real money oanda account, which will be funded by next week.
No new trades though I am very close to closing out the NZD/JPY short and going long. If so I will make sure to size the buy such that there is no chance of a margin call, (right now should all 15 trades go to SL I would be $12 short and recieve a margin call).
On the San Miguel system there was a stop loss last night on EUR/JPY short.
Overall I am very impressed with how accurate the Ichimoku is at predicting trends. Here is the accuracy for each system:
Original: 11-6 (including entire trade history)= 64% winning trades
San Miguel: 10-4= 71% winning trades
Overall Stats:
Original Ichimoku system: 3% capital risk, SL 20 pips above/below Kumo, no TS
Closed Trades: 2
Record: 0-2
Realized profit: -5%
Open Trades: 15
Record: 11-4
Unrealized profit: 27.3%
Total return to date: 22.3%/13 days =1.56% daily compound return = 28,300% annualized return
San Miguel 1%: 1% capital risk, SL at Kijun Span, TS when profit=size of SL, TS placed at entrance price +10 pip profit
Closed Trades: 3
Record: 0-3
Realized profit: -3%
Open Trades: 11
Record: 10-1
Unrealized profit: 9.7%
Total Return to date: 6.7%/8 days= .81% compounded return = 1800% annualized return
San Miguel 2%: same rules as San Miguel 1% but with 2% capital risk
Total returns to date: 13.4%/8 days= 1.58%/day compounded return= 3,000% annualized return
San Miguel 3%: Same rules as San Miguel 1% but with 3% capital risk
Total returns to date: 20.1%/8 days= 2.32%/day compounded return = 425,700% annualized return
I realize that these insane returns are based on statistically insignifigant sample sizes and very short time periods and that the short term kickass performance I am witnessing is probably based on luck, but I am tracking them regarless so we can see how different systems perform over time.
That and because, its nice, at least in the short term to imagine what those kind of numbers could mean for a real money account!
Let us hope you donât receive a margin call.
It looks like NZD/JPY short will be closing soon, at which point I will be safe from a margin call. My mistake was in making the 15 trade full size. Had I made it 40% smaller than trade 14 the lowest my available margin could drop to (assuming that every single trade goes to SL) would be $5 above the minimum margin requirement.
When I close this trade and enter a 15th I will make sure to adjust the size of the trade to make margin calls a mathematical impossibility.
Well, not much to report other than I closed out the NZD/JPY short and in the San Miguel system EUR/AUD short stopped out and I added a TS to EUR/CAD short.
Two interesting things to note:
-
Some of my profitable trades are turning, so now this will offer a great oppertunity to test the merits of a TS vs no TS on profitability.
-
The importance of Carry trades is higher than I thought. I just noticed on the original Ichimoku system that I have lost .75% of capital in daily interest charges. This comes to 17% annual interest, pretty damned high! On the San Miguel it comes to 6% or so, which makes sense since San Miguel 1% is 3X less leverage than the Original system.
It will be interesting to see how these carry trade interest charges effect overall profitability.
If any system can deliver 50% annual returns, then 18% or so interest payments are just the cost of doing business. But if the system proves less profitable then it could mean that the system canât beat investing in an S&P 500 index fund.
Of course the alternative option to giving up on Forex would be to use the Ichimoku to position trade but only do it for pairs with a profitable carry trade.
Anyway hereâs the performance of each system so far:
Original Ichimoku:
Closed Trades: 3
Win/Loss record: 0-3
Realized loss, (now including carry trade interest losses): 7.2%
Open Trades: 14
Win/Loss Record: 11-3
Unrealized Gain: 17.7%
Total Profit: 10.5%/16 days = .63% daily compounded returns= 875% annualized returns
San Miguel 1%:
Closed Trades: 4
Win/Loss record: 0-4
Realized profit: -4.2%
Open Trades: 10
Win/Loss Record: 10-0
Unrealized Profit: 7.9%
Total Profit: 3.7%/11 days= .33% daily compounded return= 233% annualized return
San Miguel Standardized: (in order to compare to the Original systemâs 3% capital risk I multiply the San Miguel performance by 3)
Total Profit: 11.1% = .96% daily return = 3,185% annualized return
What is the SanMiguel system exactly? 
I have a few different methods I use but I guess youâre referring to using the Ks as the stop? This also allows you to get back in on the trade when price comes back past the Ks.
The carry trade is a very important concept in spot forex and you should look for those pairs that give good interest eg AUDUSDâŚin fact anything against the dollar should be good as they have 0% interest.
IMO for long term trades, you should consider the cost of the carry trade and only take the trade if the move of the trend outweighs the costâŚperhaps on an intraday basis.
First to answer San Miguelâs question, the system named in his honor uses Kijun-sen to set the stop loss, has 1% capital risk and once a trade moves to profitability equal in size to the stop loss, one sets a trailing stop of (stop loss-10 pips) guaranteeing a profitable trade.
And yes, carry trading is turning out to be very important, but more on that later.
First the updates for the day. For the orginal system, closed CHF/JPY short because Kumo has turned bearish, T/K cross and Kumo Breakout indicate the trend reversed. Note due to Chikou Span not being above price curve I did not enter a long trade.
I did enter a EUR/NZD short trade though, which maxes out the original system at 14 trades.
Here is the performance so far:
Original system:
Closed Trades: 3
Win/Loss Record: 0-3
Realized profit: -10%
Open Trades: 14
Win/Loss Record: 10-4
Unrealized Profit: 23%
Total Performance: 13% gain (including carry trades)/19 days=.65% daily return= 946% annualized returns
San Miguel System:
Closed Trades: 4
Win/Loss Record: 0-4
Realized Profit: -4.2% (including carry trades)
Open Trades: 11
Win/Loss Record: 10-1
Unrealized Profit: 8.8%
Total Return: 4.6% in 14 days= .32% daily return= 219% annualized return
San Miguel Standardization: (if San Miguel were run with same capital risk as Orginal system)
Total Return: 13.8%/14 days= .93% daily= 2,809% annualized returns
Now as to the point that San Miguel made regarding carry trades: its an interesting point for sure, especially since some traders use carry trades exclusivly to make money and using the Ichimoku to due only carry trades that are trending in your favor sounds like a fine idea.
BUT I am a bit hesitant to try this for two reasons.
- The volume of trades is much lower with this system. The main benefit of the current systems is that you can have many trades open at once, so though your individual leverage/trade is low, your overall leverage is high, and so you can earn the kinds of profits that is the entire reason for Forex.
This is especially true if you were risking 1%/trade. If you can only enter a trade whoâs carry trade is in your favor then you may only have 1/3 as many trades, which combined with the lower leverage, means your returns will be adversaly effected by any losses, since the Ichimoku is a probability game.
For example, if I only carry trade based on Ichimoku, risking 1%, then I may only be in 4-5 trades at a time, with leverage of .7/trade or 3X leverage overall.
If I get stopped out on 2 of these trades then I am down 2% and may be unable to enter any new trades because the carry trade would be against me.
At the end of the year, I may only have 15 trades done, 7 or 8 of which were losses, and the gains were only at .7-.8 leverage so the actual returns are 15-20%. When accounting for short term capital gains, the net return may only be 10%, which is not competative with the S&P 500 index, which has 11% annualized returns since 1926, (this time period includes the great depression, several sharp recessions and 2 oil crises so it safe to say that 11% is the long term returns in the stock market).
- The daily interest rate for the US changes, from 0-.3% so for example had you recently entered long USD/JPY the past few days the carry trade is against you!
But if you check FX street for international interest rates you see .25% USD and .1% Yen, so you would have entered the trade and now what? Close it?
Of course it may be worth considering trying a carry trade/Ichimoku system with higher capital risk, and using the San Miguelâs TS and SL method.
The logic behind such a system is this:
The smaller SL from using the Kinjen-sen + 3% capital risk = higher leverage.
Now this would limit your max number of trades, BUT since you can only trade in the direction of a profitable carry trade, this dosnât matter so much AND you can earn a decent return on the carry trade alone. PLUS, through in the Ichimoku being used to give you a probabistic edge in terms of riding trends PLUS the TS locking in profits and minimizing losing trades and you may just have the recipe for a very nice little system that makes most of its money through the carry trade with position trading profits as icing on the cake.
OK San Miguel you have inspired me: I will track this system as well: Using FX street as my source of info for international interest rates, I will use the Ichimoku to enter carry trades that would be profitable based on the daily interest payments and use 3% capital risk and the Kijun-sen to set the SL.
Then when a trade is profitable by the amount of the SL, I set a TS equal to the size of the SL-10 pips, guaranteeing me a profitable trade.
I shall christen this system âSan Miguel CTâ.
Looks interesting.
Just a thought my friend but since I didnât invent either of these why not call them the trail by KS method and the carry trade method
Looking forward to the results as the carry trade is something that people on here donât take enough account of.
Sure thing San Miguel, KS and CT methods they are.
Well, as stated yesterday, the original system is maxed out at 16:1 leverage and 14 open trades. Now we will start to see some differentiation between that system and the KS system San Miguel pointed out.
Speaking of which, 3 trades to report. Yesterday, was the first profitable exit on the KS system, with NZD/USD short getting stopped out for a 10 pip profit, (about .1%) due to the Spike in the dollar caused by news.
Today 2 new trades, AUD/NZD short and NZD/JPY long.
That last trade was also the only trade today on the CT system, which is the new Carry Trade/Ichimoku system I am testing. Yesterday I entered the EUR/NZD short which went well over the evening as news caused that pair to tank and me to institute a 165 pip TS.
As I prepare myself for next week, when my $10 gets deposited in my Oanda account and I start trading for real money I canât help but think back on what I have already learned from my experience.
It seems that the insanely good performance of my Ichimoku based systems is based on luck, the luck of getting such a high percentage of initial trades right and those trades moving very quickly up.
Now that some of those âhot pairsâ have stalled and maybe reversing, it will be intereting to see which systems perform the best. Specifically what will be the performance difference between TS and waiting for full confirmation of a trend reversal.
Anyway hereâs the summaries of the systems so far.
Original System: 3% Capital risk, Kumo+20 SL, exit when full trend reversal
Closed Trades: 3
Win/Loss record: 0-3
Realized Profit: -10%
Open Trades: 14
Win/Loss Record: 10-4
Unrealized Profit: 25%
Total Performance: 15% gain over 21 days =.67%/day = 1,035% annualized gain.
KS system: 1% capital risk, Kinjun-sen SL, TS when profit=SL size
Closed Trades: 5
Win/Loss Record: 1-4
Realized Profit: -4.3%
Open Trades: 12
Win/Loss Record: 9-3
Unrealized Profit: 9.3%
Total Performance: 5% over 16 days= .31%/day and 204% annualized returns
Note: AUD/CAD short is 10 pips away from its TS, which the 15 min charts indicate will be hit sometime tonight.
If this occurs, this will mark the KS systemâs first notable win, a 228 pip profit that equals 1.2%.
KS standardized with Original system = exactly the same performance as the original system.
CT system: 3% capital risk, KS SL, TS when profit= SL loss
Open trades: 2
Win/Loss Record: 1-1
Unrealized Profit: 4%
Total Performance: 4% in 1 day = 1,650,000,000% annualized interest*
- note that this performance is entirely meaningless, as the number of days increase this perfromance will drop exponentially, I only include it because the number is awe inspiring in its scale and its fun to imagine what we would do with all that money.