Live Forex Charts - FXStreet

Just want to share my experience doing Norbit's Gambit, for anyone that may want to learn on my experience, or tell me all the ways in which I am wrong.
I've posted a question earlier, asking how to make sure the market didn't swing in the wrong direction. And yes, I understand it's called a "gambit" for a reason.
=== TL;DR === - Wanted to save $65 on currency conversion while buying CAD $3,500 worth of US stock - Overestimated forex fee, underestimated NG cost - Murphy happened... - Lost potential gains of CAD $185.80 while waiting - Is the gambit ever worth it?
=== The goal === - On August 25th, use about CAD $3,500 to buy some US Netflix (NFLX) shares
=== The method === - Buy DLR in CAD - Call brokerage to journal over to US side - Sell DLR.U in USD - Buy NFLX with USD
In a lot of my research, I've been told the "fee" when going through the brokerage (Questrade in my case) is 2%. That's not too accurate.
=== The brokerage Forex rate/fee === So I assumed 2% of $3,500 would be $70.
However it's really an addition of 199 basis points to the exchange rate. I believe they use the closing exchange rate on the date of the transaction.
So if the exchange was 1.2500, it's 1.2699, and $CAD 1000 nets $USD 787.46 ($USD 12.54 short of ideal $USD 800). If we convert 12.54 to CAD, it's $CAD 15.68 or 1.568% on the original $CAD 1000.
But if the exchange is 1.2200, it's 1.2399, and $CAD 1000 nets $USD 806.51 ($USD 13.16 short of ideal $USD 819.67). If we convert 13.16 to CAD, it's $CAD 16.06 of 1.606% of the original $CAD 1000
The "conversion fee" is dependent on the exchange rate, but I can't figure out a quick direct way to calculate what the "fee" would be in source currency %. The closest guesstimate is to divide the basis points by the current exchange rate, as below:

• In first example, 0.0199 / 1.25 = 0.0159 or 1.59%
• In second example, 0.0199 / 1.22 = 0.0163 or 1.63%

On August 25th, the exchange rate was 1.2483. So the above guesstimate for the Forex "fee" would be $55.80. Or more accurately $54.91, close enough. Still, that's 21.5% lower than the broad *$70** estimate earlier.*
=== The NG effective fee === Here once again, the common consensus is that the NG's fee is just a cost of buying and selling an ETF. So, with QT's free ETF purchases, the guesstimate is just about $5 for the selling commission.
When I bought DLR ETF on August 25th, I got it for CAD $12.42 per share (that's not the day's close price, but what I actually paid). the US side of DLR.U is always US $9.93. The gives the DLR Exchange Rate of 1.2508... already different from the ideal 1.2483 (I suppose that's how the ETF makes money)

• CAD $3,502.44 buys 282 whole shares of DLR (I will address the extra $2.44 later)
• No buying comm, but ECN of CAD $0.0035 * 282 = $0.99
• Selling comm and ECN: US $0.99 + $4.95 = $5.94

So, it would appear the cost/fee of NG is just CAD $0.99 + US $5.94 (CAD $8.40)... but not quite. Let's look as the actual US dollar amounts. Since we can only buy whole shares of DLR + ECN fee, from this point I am converting CAD $3,502.44 + $0.99 = $3,503.43

• Ideally, with exchange of 1.2483, CAD $3,503.43 nets US $2,806.56
• I paid CAD $3,503.43 to buy 282 DLR shares
• I sold DLR @9.93 for US $2,800.26
• Less US $5.94 comm+ENC leaves me with actual cash of US $2,794.32

The difference between the ideal US $amount and what I am left with is the "fee" for doing NG. US $12.24, or CAD $15.28. This sets the NG's Effective Exchange Rate as 1.2538 or a "NG's fee" of 0.44%. Please note that this rate/fee is *based on the converted amount*: the more you convert, the less/cheaper it is.
So comparing the actual cost of NG vs QT's auto conversion and the guesstimates is quite different - Actual $15.28 vs $54.91 ($39.63 spread) - Guesstimate $5 vs $70 ($65 spread)
=== The Time factor === Now that I have US $2,794.32, let's buy some NFLX. The date now is Sept 5th (yes, the NG completed earlier, but I am human... also QT didn't call me when the journaling was completed so I got sidetracked).
Before doing the NG, I calculated that NFLX dropped about 2.02% in 10 days. It could also go up by same amount. That 2.02% rise on CAD $3,500 value would be a gain of $70.7... comparable to $65 guesstimate loss of doing currency auto-conversion through QT (QTAC). So in my analysis, at worth case it would be a wash, in best case I save some money on NG. ... We already know now that the actual cost of QTAC is much less ....

• On September 5th, NFLX traded (at time of my purchase) at $178.79
• That's a 6.74% increase of August 25th's price of $167.50 (at the time of me recording the price)
• US $2,794.32 buys 15 whole shares for $2,681.85 + $4.95 Comm = US $2,686.80 spent. [Weirdly enough I had no US ECN fees in spite of doing odd lot market orders]
• This leaves US $107.52 in cash
• Total US Portfolio value (cash + NFLX value) $107.52 + $2,681.85 = $2,789.37

This was the end result of my NG: spent CAD $3,503.43 and 11 days later I have a US stock+cash Portfolio of value US $2,789.37
But considering the increase in NFLX stock over those 11 days, what if I would have just went with Questrade's Auto-conversion (QTAC) route?

• QTAC exchange rate on August 25th, 1.2483 + 0.0199 = 1.2682
• CAD $3,503.43 will net US $2,762.52
• That would buy 16 whole NFLX shares @167.50 for US $2,680 + $4.95 comm
• Total spent US $2,684.95 and left with $77.57 cash
• Cash + NFLX Portfolio value US $2,757.57

In 11 days, the value of NFLX is @178.79 - Current value of 16 NFLX shares = US $2,860.64 - Plus cash US $77.57 for a Total Portfolio Value of US $2,938.21
If, on August 25th, I would have auto converted currency with QT and bought NFLX, I would have a portfolio value of US $2,938.21 on September 5th.
Instead, starting the Norbit's Gambit on August 25th, I bought NFLX on Sept 5th and have a portfolio value of US $2,789.37
Instead of saving a guesstimate of CAD $65, I have lost potential gains of US $148.84 or CAD $185.80
=== Murphy... or Loonie... whatever === In all of above, I tried to keep the currency fluctuations isolated. So apart from initial conversion on August 25th, all my future (September 5th) portfolio values were in USD. But as Murphy would have it, the BoC rate announcement made the loonie stronger in between my NG.

• On September 5th, the ideal exchange rate jumped to 1.2235
• DLDLR.U had a somewhat delayed reaction, and was almost same at 1.2236

Based on above, doing QT auto-conversion - On August 25th, nets US $2,762.52 - On September 5th, nets US $2,817.62 I would have got US $55.10 more just by doing the QTAC later.
If I would have bought NFLX on September 5th after doing QTAC - 15 whole shares @178.79 + $4.95 comm leaves me with US $130.82 in cash - Total NFLX + Cash portfolio value of US $2,812.67
That's still US $23.30 more than doing the NG, although still less than just buying NFLX outright on August 25th and letting it grow.
=== Final conclusion === Smaller amount (CAD $3,500) for NG for a stock purchase that could/did swing 6% was not worth it. Loonie getting stronger also made the whole exercise fruitless, but even eliminating the currency fluctuation, the growth of the stock outperformed the savings of NG.
Playing with my numbers, assuming the currency fluctuation is fixed, for CAD $3,500, doing Norbit's Gambit vs Questrade's Auto-Conversion is breaking even when the stock appreciation is no more than 1.18% during the time it takes to complete NG (11 days in my case). Even if you are more punctual and can complete it in 5 days, you still need to make sure the stock doesn't appreciate more than 1.18% in 5 days.
By comparison, if converting CAD $10,000, it's break even if stock rises 1.27% during that time. When doing CAD $50,000 then 1.31%
Hmm.... so even at high amounts of CAD $50,000 the tolerance to stock fluctuation is pretty low. So is it worth it?
Ultimately I've:

• Overestimated the Forex fee
• Underestimated how much NFLX could swing
• Got distracted to do it in more timely manner
• Also ... loonie...

For anyone that cares, here is public Google Sheets docs to verify my math (File -> Make a copy, in order to edit values) https://docs.google.com/spreadsheets/d/12nH8Q_0cKFtFTW_yomASHtpLXN6oYuTgkXeTNAMkQGY/edit?usp=sharing

Looks like yesterday I didn't give myself enough leeway with my title because I don't think anybody expected what happened yesterday to happen. There was a double digit decline in the RUB's value and a huge emergency rate hike late at night by the RCB.
The latest numbers from the Cenbank are in and they just spent Over $2 billion propping up the ruble
If you want to follow:spin0 has linked us these nice sites with good charts: http://forex.tradingcharts.com/chart/US%20Dollar_Russian%20Ruble.html
http://www.xe.com/?c=RUB&r=
If you have another good site to monitor the situation (like zenrus, post in the comments! I'm going to try and make this post a little easier to follow too.

Time (EST)	USD/RUB	±RUB/USD	±%
7:02am	75.1710	+9.5880	+14.6200%
8:06am	72.4535	+6.8705	+10.4800%
9:03am	72.9235	+7.3405	+11.1900%
11:34am	68.3180	+2.7350	+4.1700%
1:00pm	72.2495	+ 6.6665	+10.1600%
2:47pm	67.5005	+ 1.9175	+2.9200%

Right now I use Forex at TradingCharts.com. It is pretty good, but I was wondering if there was something I hadn't seen yet.