I think its rediculous that traders could think that the USD rally is over, and hence bet against the USD (as indicated with the SSI numbers). Looking at fundamentals, as well as technicals, it seems the USD is poised to continue its rally for some time. I put up a chart in the Fibonnaci trading forum that shows that the GBPUSD just broke a very crucial fib line, and that's just one example. As many see Europe and Great Britain entering a turbulent time, I encourage people to stop trying to predict when the trends will end, and instead, trade WITH the trend! Go long on the USD!

One of my most frequent pieces of advice to anyone who asks for it is to trade with trends as they make up 90 percent of the market (up, down and sideways); but for some reason retail traders have a natural attraction to trying to pick major reversals. That is essentially the gambling play. Trying to call a major turn with lots of retracement and a tight stop (because they think there is no risk as it's reversing). This is a behavior that will never fade unless cultures change. Of course, in this inefficiency, there are trading opportunities...

__________________
John Kicklighter is the author of Dynamic Carry Trade Basket, Watch What The Fed Watches, and Forex Trading Weekly Forecast on DailyFX.com

David,

I've followed the retail holdings data on this site and two others for over a year, and the general conclusion I've come to is that retail in aggregate is wrong until they're not. What I mean by that is demonstrated in today's action. Going into today, short USD weightings in various pairs ranged from 55-80% of net positioning. Today those pairs are up nearly a full percent. So while retail had been "wrong" for some time, today they're "right". And from my perspective, there was no determinable change in positioning recently that would have generated a buy/sell signal going into today's USD selloff.

I think the overall premise that retail is always wrong, i.e. the market always moves in the opposite direction of the net retail position, is certainly open to debate.

On a separate note, can you confirm whether or not your positioning data is weighted by position size? One site in particular makes no distinction, why, I have no idea (weighting equally a 1-minilot position with a 100-lot position is pretty much invalidating the data entirely). I'm hoping this is the reason why your data sometimes stands in stark contrast with the other holdings reports (yours being more accurate hopefully).

I'm not David, but I have some opinions for these questions.

It it certainly not true that retail traders are 'always' wrong. In fact, they do quite well in range conditions and we have actually seen this in the market conditions from March until July. This strong performance however further highlights their shortfalls. They look for simple technicals that are blatantly obvious because it fits unreasonable risk/reward scenarios that are usually applied. More influential though is the habit for letting losses run and taking profits far too early.

So, while traders may be right in quiet markets, even then they quickly fight moves in ranges and put positioning on the wrong side of the move. But the big shifts - trend changes, major breakouts - they often fight. We can see this in the longer term for the EURUSD and GBPUSD breakdowns.

Also, just like any indicator, it isn't wise to use an indicator as an absolute signal. The SSI is better to see longer term trends in positioning and mark extremes as well as significant shifts for projecting spot action.

We have a weighted and unweighted version of the positioning data, but I don't remember which we use for the SSI report. However, I would argue that an unweighted version is best. Consider, FXCM has large clients that are institutional, professional, systems, etc. Therefore, if you are trying to get a reading for what retail is doing, you would get a big skew from these deep capital pools that have a different style and mentality. Equal weighting for all gives a better feeling for the crowd - which is what we want to see rather than see where the bigger people are trading (which can already be derived from COT data).

__________________
John Kicklighter is the author of Dynamic Carry Trade Basket, Watch What The Fed Watches, and Forex Trading Weekly Forecast on DailyFX.com

As John was so enthusiastic to say in the post above, the retail investor is certainly not always not always wrong. But the key here is that the retail investor typically goes for smaller moves, remains highly leveraged, and tends to fade price trends.

Using the SSI as a contrarian indicator would've proven very unprofitable in the month of June, for instance, because currencies were largely rangebound and the retail investor remained, on aggregate, profitable. (I know this fact very well because of returns in relevant SSI-based strategies.) Anecdotally, professional traders suffered big time in June because of the choppy trading ranges. That brings me to my next point:

The SSI is a good trading tool despite the fact that it will be absolutely wrong in certain circumstances. In fact, it may very well be true that it is more often wrong than right (I don't think this is true, however.). But SSI-based trading signals will prove quite accurate in price trends, and as professionals will tell you, this is where the real money is to be made. If a trader is comfortable remaining unprofitable for months at a time in rangebound markets, only to hit it big in several weeks of strong trend moves, then the SSI can be a very helpful tool in anticipating these big moves.

[/quote]

After extensive research, we found that it is most helpful to not weight by position size in our SSI indicator. John offered his opinion above, and I mostly agree. We assign each trader 1 vote: long or short. Regardless of whether that trader is in the market for 10k or 2.0m, he or she has one vote. Though this seems counterintuitive at first glance, our own research tells us that doing this provides more accurate trading signals. I would argue that our data seems more "accurate" than most because of our large sample size. That is to say, we have access to one of the largest pools of traders of any FX firm.

Thanks for the great questions. Hope I answered them well, and as always I'm willing to answer any further questions on the SSI.

__________________
David Rodriguez is the author of Forex Trading Signals and Forex Trading Weekly Forecast on DailyFX.com.

John,

Could you please elaborate more on this? What types of "simple technicals" are they looking for, and how do these differ from the various types of technical analysis reviews provided by your site's analysts? I would think they would be following the same or similar technical signals and strategies your analysts provide (which, with some exceptions, provide profitable signals).

Agreed 100%. Clearly evident by how quickly the "right" positioning dives off as soon as the trend starts running in their direction.

Good points, agreed.

If you or David could confirm weighted vs unweighted for the SSI report, that would be great, thanks. (edit - David's reply, thx). I think offering both sets of data would be the best thing to do (please!).

While I agree that an overall crowd picture (unweighted) has its benefits, tracking the weighted/large positions in real-time, could be very advantageous. The reason being, the COT is a single snapshot that is nearly a week delayed when it is released. It's usefulness in terms of short-term trading is limited by this issue. It's better for long-term position trading.

A real-time weighted snapshot has significant benefits. Let's take this one step further and use today's market action as an example. Say there are two versions of the SSI released, weighted and unweighted. Looking at the end-of-day reports for both sets of data could provide rather valuable clues. Hypothetically, let's say that today's data shows a reversal in USD positioning from short to long in the weighted category, while remaining short-USD (or perhaps increasing short) in the unweighted category. In essence, institutions exiting their profitable short-USD positions after the 1% jump, and buying long-USD positions from retail shorts, who are typically late to enter upon seeing this "trend breakout". Or, the complete inverse of that scenario could occur. Either way, that data could be rather advantageous to track, would you agree?

(edit - sorry for various corrections, busy day)

David,

All very good points, thanks for sharing your thoughts.

Interesting, thanks. Per my last post, what are your thoughts on the more short-term implications of both data sets? (weighted and unweighted) This might be more difficult to backtest empirically, but I think it would prove beneficial to have the data accessible for review, especially at major market inflection points (the USD acceleration in early August, potentially the action today, etc).

By simple technicals, I do not mean that any indicator is inferior to another (in fact, I actually know people who are incredibly consistent and profitable with two exponential moving averages). What I mean is that many of those new to this fast paced market or trading in general don't know how to use technical and fundamental analysis properly; but more importantly they don't know how to set up a sound trading strategy and follow it.

If I gave a group of brand new traders 10 trades that are laid out plainly and end with a 100% accuracy rate, there is still a good chance that the majority will lose money. This is because they don't know how to manage a position, the will try to cut corners on stops, push the limits on targets, have late entry, use too much leverage etc.

Good theory, but there are a few things to take into account. First, FXCM is the largest retail broker in the world and using it as a signal as such is sound because of the sample size. Using a weighting system, you wouldn't be measuring a bias anymore, rather you would be skewing towards though institutions, money managers, systems that are trying to take advantage of the same inefficiencies found in retail traders - so you would in fact be watering down the data.

Also, the sample size of the non-retail group is far smaller than that of the regular speculator, so we no longer have a statistically strong reading.

Great questions btw!

__________________
John Kicklighter is the author of Dynamic Carry Trade Basket, Watch What The Fed Watches, and Forex Trading Weekly Forecast on DailyFX.com

When I was updating the SSI this afternoon, I noticed that retail traders are holding little bias despite what could be the reactivation of some major trends (like the new 6-month low in EURUSD yet the ratio stands at 1.27 with only 56% percent of the pool long). I updated the spreadsheet below. Take a look at it.

Attached Images

 

__________________
John Kicklighter is the author of Dynamic Carry Trade Basket, Watch What The Fed Watches, and Forex Trading Weekly Forecast on DailyFX.com

I have 3 brokers I am working with and all three show a different close for USD/JPY. Can any one tell me why that is, what causes a different close price and any one time.

FX_Solution 107.35
VT-Trader cmsfx.com shows 107.18
ODL Securities 107.73

How can this be? I also notice that before the market opens these numbers start changing on Sunday.

Same with all the pairs.
EUR/USD for instance
ODL - 1.4267
CMSFX - 1.4234
FX Solutions 1.4235


Which is right? Which broker is doing something odd or wrong?

The reason for the price discrepancies is because some brokers, for whatever misguided reason, consider the close of FX trading on Friday to be anywhere from 3:00 - 4:30pm EST. Similar brokers also open either early or late on Sunday. The interbank market (EBS, Currenex, HS FXi, etc) trades from Sunday 5pm EST to Friday 5pm EST. All brokers should adhere to those hours for the benefit and safety of their clients. Unfortunately, very few dealing desk brokers remain open until 5pm on Fridays..

This past Friday highlights this importance. The Fannie Mae/Freddie Mac bailout rumors began hitting the wires around 4:30pm, causing a spike against the dollar, as well as a larger spike in the yen pairs. ODL's closing prices for both EUR/USD and USD/JPY reflect the correct 5pm prices. FX Solutions closes early around 4:30 (thus their clients missed the FMae/FMac news & price movements). CMS also closed around a similar time. I'd avoid both brokers, and any others that don't adhere to the interbank trading hours.

Hey all,

I just posted my weekly Speculative Sentiment Indicator report here: Forex Trading Indicator Accurately Forecasts Euro/US Dollar Declines - What's Next?

This is the perfect place to ask me any and all questions about the report and the SSI itself.

__________________
David Rodriguez is the author of Forex Trading Signals and Forex Trading Weekly Forecast on DailyFX.com.

I have an open position of EURJYP of 160000 (buy) @150.290 my loss is 7000USD how can i close this deal with the minimum damage? Thanks

Thursday, 09 October 2008 14:31:55 GMT
Written by David Rodriguez, Quantitative Analyst
Full Article

• EURUSD – Forex Trading Indicator Forecasts Euro May Gain Despite ECB Rate Cuts
• USDJPY – Japanese Yen Predicted to Strengthen Further Through Financial Crisis
• USDCHF – Forex Trading Signals Buy the USDCHF on SSI-based Forecasts
• GBPUSD – Traders Indecisively Buy the US Dollar Against the British Pound
• USDCAD – Canadian Dollar Tumbles as Contrarian Indicator Shows Weak Bias

While the SSI is available once a week on DailyFX.com, you can receive SSI readings twice a day in DailyFX Plus Forex Intraday Trading Signals. The SSI sought a EURUSD rally since 1.26 and was signaling a reversal around 1.60.



How to We Interpret the SSI? The FXCM SSI is based on proprietary customer flow information and is designed to recognize price trend breaks and reversals in the four most popularly traded currency pairs. The absolute number of the ratio itself represents the amount by which longs exceed shorts or vice versa. For example if the EURUSD ratio is 2.55, long customer orders exceed short orders by a ratio of 2.55 to 1. Conceptually similar to contrarian analyses using the CFTC IMM open position data or COT Report, the SSI provides an alternative approach that is both more timely and accurate in forecasting currency price movement. The SSI is a contrarian indicator that tells you how the market is weighted and where the trend may head. More long positions don't necessary suggest more confidence in the direction of the current trend. In general, when traders start having adverse movements against their position, many tend to increase the size of their position with the purpose to average down their entry price in one last attempt to recover from previous losses. However, the higher the number of short orders in a bull market the more dangerous is to take additional shorts because many of those traders who just entered the markets are also leaving their protective stop losses just above the current price action.

David/John,

I sent Jamie Saettele a PM a week ago & waiting to here back, also noticed he's not doing the daily review this week. Is he just out for the week? thanks