My multi-currency trade journal

Update as of 9:00pm EST, 9/2/14
Current month start date: 8/19/14
Current month end date: 9/16/14
Month 4, Day 12

Account balance (8/19/14 AM): $213.60
Account balance (9/2/14): $216.86
%gain = 1.52%
Floating P/L = -$42.62
Net Asset Value = $174.23
% Floating DD = 19.7%
Positions open: 35
Average No. units per position = 52
Margin used = $46.48
Margin Available = $127.75
% Margin used = 36%
RISK = ~$2.60 per 500 pips draw down per position (= ~$91 for all open positions)
%RISK = $91/$216.86= 42%

Update as of 9:00pm EST, 9/3/14
Current month start date: 8/19/14
Current month end date: 9/16/14
Month 4, Day 13

Account balance (8/19/14 AM): $213.60
Account balance (9/3/14): $216.86
%gain = 1.52%
Floating P/L = -$46.15
Net Asset Value = $170.67
% Floating DD = 21.2%
Positions open: 35
Average No. units per position = 52
Margin used = $46.53
Margin Available = $124.14
% Margin used = 37%
RISK = ~$2.60 per 500 pips draw down per position (= ~$91 for all open positions)
%RISK = $91/$216.86= 42%

Update as of 9:00pm EST, 9/4/14
Current month start date: 8/19/14
Current month end date: 9/16/14
Month 4, Day 14

Account balance (8/19/14 AM): $213.60
Account balance (9/4/14): $216.84
%gain = 1.52%
Floating P/L = -$67.87
Net Asset Value = $148.98
% Floating DD = 31.2%
Positions open: 41
Average No. units per position = 54
Margin used = $55.95
Margin Available = $93.03
% Margin used = 60%
RISK = ~$2.70 per 500 pips draw down per position (= ~$111 for all open positions)
%RISK = $111/$216.84= 51%

Update as of 5:00 pm EST, 9/5/14
Current month start date: 8/19/14
Current month end date: 9/16/14
Month 4, Day 15

Account balance (8/19/14 AM): $213.60
Account balance (9/5/14): $217.09
%gain = 1.63%
Floating P/L = -$68.10
Net Asset Value = $148.99
% Floating DD = 31.3%
Positions open: 41
Average No. units per position = 54
Margin used = $55.97
Margin Available = $93.02
% Margin used = 60%
RISK = ~$2.70 per 500 pips draw down per position (= ~$111 for all open positions)
%RISK = $111/$217.09 = 51%

I've uploaded my "Percent Directional Similarity" spreadsheet for all currency pairs that I trade (7 majors and 21 crosses). As briefly explained in post 155 of this thread, the algorithm first assigns a binary score ("1" or "0") to the week-to-week directional similarity of exchange rate movement for each currency pair relative to every other currency pair in my watchlist, it then adds up the binary score over a 52 week look-back period, and then the "percent directional similarity" is calculated for each currency pair comparison based on the 52 week look-back period. Basically, I wanted to answer this question: How many times during the past 52-weeks did the week-to-week exchange rate of one currency pair move in the same direction as another currency pair? So for instance, as of August 31, 2014, the 52-week Directional Similarity of EURUSD vs. USDCHF = 6%. This means that the exchange rates of the EURUSD and USDCHF moved in the same direction only 3 weeks out of the previous 52 weeks (August 30, 2014 - August 31, 2014). Compare this to EURJPY vs. CHFJPY; their 52-week Directional Similarity = 87%, meaning that both exchange rates moved in the same direction 45 weeks out of the previous 52 weeks. Also included in the spreadsheet for each pair comparison is the yearly "Average Directional Similarity" calculated over the time period of January 2008-August 31, 2014 (N=7 yearly time periods, although the current year is not yet complete, so the column marked "2014" is actually the rolling 52 week period of August 30, 2013 -August 31 2014 ), as well as the yearly standard deviation- which gives an indication of the volatility of the "sameness" of the directional movement on a year-to-year basis. Obviously, as time passes and more data is collected, the Average & standard deviation will be a better reflection of the directional similarity over time...

The calculation of the binary week-to-week directional similarity of two currency pairs is very simple: for both pairs, calculate their most recent weekly close minus their weekly close from the previous week, and then compare these two weekly pip differentials. If both weekly pip differentials are positive, or if both are negative, or if both are zero (this never has happened), then the comparison is scored as a "1". Otherwise ALL other weekly pip differential comparisons are scored as "0", i.e. If one weekly pip differential is positive and the other is either zero or negative = score "0", or if one weekly pip differential is negative and the other is either zero or positive = score "0", or if one weekly pip differential is zero and the other is either positive or negative = score "0". These weekly binary scores are then added up for the previous 52 weeks and the Percent Directional Similarity is calculated (i.e. 52-week Sum of the Binary Score/52). Note that the "move in the same direction" can be as little as 1 pip. This is not “correlation” since the statistical parameters used for correlation are much more comprehensive than the calculations used in my simple algorithm. I picked a rolling 52 week period as a means to assess the directional similarity because I wanted to know the week-to-week variation within this time frame since most of my trades are long-term. However, the calculation can be used on any time frame.

For me, the information in this spreadsheet is used to determine which currency pairs should be hedged (positive or negative) based on their historical similarity of price direction. I hope that this information may be useful to you as well...

Dave

Dave,

I've been reading your journal with great interest since stumbling upon it a couple of days ago. Thank you for sharing your analysis; it indeed is quite interesting. I'm still wrapping my head around some of the more elaborate posts of yours; especially the indices.

For example, since you trade with a "buy all" mentality and EURUSD and USDCHF rarely move in the same direction, if you went long EURUSD, you go long USDCHF? On the contrary, since EURJPY and CHFJPY often move in the same direction, if you went long EURJPY, you would have to go short CHFJPY and hence you would not use CHFJPY to pseudo-hedge EURJPY?

Thanks for provoking thought.

Neil

Hi Neil,

When I initially ran my back tests, I used a "buy all" approach for all currency pairs as mentioned in my first post... and the results were reasonably good. This approach allowed me to set up pseudo-hedges without too much thinking... in other words, I didn't necessarily hedge specific currency pairs. There were always currency pairs moving up and pairs moving down, so I let the market dictate when to close positions; I would close the largest profitable position(s) and balance it/these by closing a losing position(s) when my minimal target %gain was reached. With that said, I managed all of the open positions as one big inventory, closing profitable & losing positions when needed. As I stated above, this approach worked reasonably well in back testing, but as you can tell by my live updates, hasn't worked so well in live trading. The major problem with this approach is that its easy to sit at my laptop at nights & weekends and run the back tests and make the necessary inventory decisions when need be, but it's harder in real time when I'm at work and don't have the ability to constantly monitor my positions.

So... what to do? Well.... I've always been interested in "un-directional" trading, and thus want to develop an approach that is more precise in managing "true" pseudo-hedges, but also allows me to have the flexibility I need to manage these positions (i.e. doesn't need constant monitoring). I'm still working out the details on the latter... But, this did lead me to develop my "All Pairs" Composite Index as well as the composite indices for each of the major pairs; when I mentioned these indices in my first post I hadn't yet incorporated their use into my trading decision trees. I then worked out the simple algorithm I use to calculate the "Directional Similarity" data as well as my "Relative Volatility" calculator, and have slowly incorporated these as well. On page 8 of this thread, I wrote several posts on how I am using the indices and the Directional Similarity information. As you pointed out, the EURUSD vs. USDCHF is a "natural" buy/buy hedge since these two pairs rarely move in the same direction, same as the EURCAD vs. CADCHF, GBPAUD vs. AUDUSD, EURGBP vs. GBPCHF etc.... On the flip side, CHFJPY vs. EURJPY or GBPJPY, GBPAUD vs. GBPNZD, NZDCAD vs. NZDUSD have a high tendency (>70%) to move in the same direction, so these would be "natural" buy/sell pseudo-hedges. I'm currently working on the trading approach to take advantage of these natural pseudo-hedges. This also means that I have to take into account their volatility differences as well- hence the use of the Relative Volatility calculator to determine how many units of each currency pair in a pseudo-hedge that will need to be opened (either buy/buy or buy/sell) in order to, volatility-wise, balance the hedge. As you can see, this is still a work in progress (and probably always will be), and I will be using my live account to test these ideas...

Dave

Dave,

From what you've posted, your results weren't bad at all. Correct me if I'm wrong but you were still reaching your 4-5% targets weren't you? Also, when you say you let the market dictate when you closed positions what exactly do you mean? I'm assuming you mean you waited until your profitable trades were at 1%. And this did not always work out during as you weren't available to monitor your trades all day?

Have you considered leveraging the programmable features of MT4 to create a sort of "watch dog" to alert you when your profits are at 1%? It seems that could be a solution especially if 7 years of back tests have proven profitable.

Indeed. It would be very interesting to see how these hedges actually fared in live trading.

Update as of 11:00 pm EST, 9/9/14
Current month start date: 8/19/14
Current month end date: 9/16/14
Month 4, Day 18

Account balance (8/19/14 AM): $213.60
Account balance (9/9/14): $219.97
%gain = 3.0%
Floating P/L = -$64.31
Net Asset Value = $155.66
% Floating DD = 29.2%
Positions open: 35
Average No. units per position = 53
Margin used = $47.98
Margin Available = $107.68
% Margin used = 45%
RISK = ~$2.65 pips draw down per position (= ~$92.75 for all open positions)
%RISK = $92.75/$219.97 = 42%

Neil,
Yes, during my back testing, I closed out combinations of profitable trades + losing trades when the %gain for those combinations achieved my target of 1%. Again, easy to do when sitting at my laptop back testing the data. In "Live" trading... I would be up 0.4%- 0.8% on a combination but then be in situations at work where I couldn't access my phone to monitor the trades.. and you won't believe how quickly those gains turned into losses.. especially when the overall market (ALL PAIRS Composite Index) is moving lower. This can't be rectified by setting predetermined stop losses/take profit targets because I never know at what point (pip wise) when to terminate a combination of profitable/losing trades (or what combination to terminate) other than seeing that a combination has yielded the 1% target gain of the account balance. Hope this makes sense. Regarding trading alerts, I don't use MT4 and even if I were to use the MT4 platform from Oanda, I still wouldn't be able to access my phone when needed to close positions. As I noted in my previous post, I'm still working out ways to manage this issue in live trading...

Dave

Update as of 8:30 pm EST, 9/10/14
Current month start date: 8/19/14
Current month end date: 9/16/14
Month 4, Day 19

Account balance (8/19/14 AM): $213.60
Account balance (9/10/14): $220.14
%gain = 3.1%
Floating P/L = -$60.46
Net Asset Value = $159.68
% Floating DD = 27.4%
Positions open: 34
Average No. units per position = 53
Margin used = $46.08
Margin Available = $113.62
% Margin used = 41%
RISK = ~$2.65 pips draw down per position (= ~$90.10 for all open positions)
%RISK = $90.10/$220.14 = 41%

Dave,

I hear you. Just monitoring trades from my phone, I've seen this happen many a time. You'll be up and in a moments notice you'll be down. It's truly amazing how fast the market moves sometimes.

Neil

Update as of 8:30 pm EST, 9/11/14
Current month start date: 8/19/14
Current month end date: 9/16/14
Month 4, Day 20

Account balance (8/19/14 AM): $213.60
Account balance (9/10/14): $221.86
Current month's %gain = 3.9%
Floating P/L = -$51.53
Net Asset Value = $170.32
% Floating DD = 23.2%
Positions open: 33
Average No. units per position = 53
Margin used = $44.02
Margin Available = $126.30
% Margin used = 35%
RISK = ~$2.65 pips draw down per position (= ~$87.50 for all open positions)
%RISK = $87.50/$221.86 = 39%

Update as of 5 pm EST, 9/12/14
Current month start date: 8/19/14
Current month end date: 9/16/14
Month 4, Day 21

Account balance (8/19/14): $213.60
Account balance (9/12/14): $223.16
Current month's %gain = 4.5%
Floating P/L = -$46.13
Net Asset Value = $177.03
% Floating DD = 20.7%
Positions open: 29
Average No. units per position = 52
Margin used = $38.43
Margin Available = $138.63
% Margin used = 28%
RISK = ~$2.60 pips draw down per position (= ~$75.50 for all open positions)
%RISK = $75.50/$223.16= 34%

Updated money flow indices for EUR, GBP, and CHF.

5-day trend for EUR, GBP, and CHF indicates recent short-term money flow into these currencies (remember, each index is built with the EUR, GBP, or CHF as the quote currency. Therefore an upward sloping trendline indicates recent short-term weakness, i.e. money flowing out of the quote currency. A downward sloping trendline indicates recent short-term strength, i.e. money flowing into the quote currency). On the left side of each index window are the constituent currency pairs for the index.

Updated money flow indices for AUD, NZD, and CAD.

5-day trend for each index is up (index weakness), indicating recent short-term money flow out of these quote currencies.

Updated money flow indices for USD and JPY. The 5-day trendline for the USD is sloping downwards indicating recent short-term money flow into the USD. On the other hand, the 5-day trend for the JPY is sloping upwards indicating recent short-term money flow out of the JPY.

Updated WEEKLY composite indices for AUD, NZD, CAD.
Slope of 12-week trendline:

AUD = DOWN = AUD strength = longer-term money flow was into the AUD
NZD = UP = NZD weakness = longer-term money flow was out of the NZD
CAD = DOWN = CAD strength = longer-term money flow was into the CAD

Updated WEEKLY composite indices for EUR, GBP, CHF.
Slope of 12-week trendline:

EUR = UP = EUR weakness = longer-term money flow was out of the EUR
GBP = UP = GBP weakness = longer-term money flow was out of the GBP
CHF = UP = CHF weakness = longer-term money flow was out of the CHF

Updated WEEKLY composite indices for USD and JPY
Slope of the 12-week trendline suggests:

USD = DOWN = USD strength = longer-term money flow into the USD
JPY = UP = JPY weakness = longer-term money flow out of the JPY