PM Daily Market Commentary - 5/9/2017

By davefairtex on Wed, May 10, 2017 - 2:01am


Gold fell -4.90 to 1221.30 on very heavy volume, while silver dropped -0.05 to 16.18 on moderate volume. It was a second strong day for the buck, and a weak day for the metals.

Gold made a new low in Asia to 1214.30, and looked as though it was going to close near the lows, until, in the last 30 minutes of trading in the US, the equity market sold off sharply, while TLT, gold, and silver all rallied equally sharply. Various currency pairs also moved strongly, with JPY making the largest move: 0.40 points in about 10 minutes. Supposedly this was triggered by an interview conducted by Sky TV with the North Korean Ambassador who threatened to turn US strategic assets into ashes if the US executed a military strike on his country.  Both gold and silver held their gains into the close.

The candle print for gold was a spinning top which the code found to be somewhat bullish. Gold's RSI-7 is now 16; in USD, gold is increasingly oversold. Viewed in Euros, gold has been chopping sideways for the past few days; most of gold's move today was a currency effect.

Open interest at COMEX for GC fell -1,171 contracts.

Rate rise chances (June 2017) fell to 83%.

Silver also fell for most of the day, making a new low to 16.06 before finding buyers right at end of day in the US. Candle code felt the spinning top candle was neutral. No reversal bar today for silver. Silver is approaching support at round number 16. Silver has fallen for 16 of the last 17 days, and the RSI7 for silver is now 6.10, which is extremely oversold.

Open interest at COMEX for SI rose +3,821 contracts.

The gold/silver ratio fell -0.07 to 75.46.

Miners gapped down at the open, but moved steadily higher during the day. By the close, GDX was off just -0.09% on light volume, while GDXJ was down -0.33% on very light volume. Candle print for GDX was a short white candle, which the code found to be relatively bullish. GDXJ's candle was a short white/NR7, which the code thought was slightly bullish. GDX:$GOLD (actually, GDX:GLD) moved higher today, moving more conclusively above its 9 EMA. It appears to have marked a low. The GDXJ:GDX ratio dropped slightly, but still appears to be chopping sideways.

Platinum was hit hard, dropping -1.50% and printing a very bearish swing high candle pattern. Palladium dropped -1.56%, printing its own very bearish swing high-like pattern, while copper rose +0.08% printing a doji which the code felt was bearish too.  The "other metals" did not look good today, especially platinum which appears to be headed for a re-test of the lows.

The buck rallied strongly for a second day, moving up +0.60 to 99.45, printing a two candle swing low that the code felt was extremely bullish. The buck is now convincingly above its 9 EMA, as well as the 200 MA. Is this about the Fed and its plans to reduce the balance sheet? Its certainly not about Macron's victory in France.  Or maybe it is in a way - just as a sell-the-news event: the Euro fell -0.46% today, and is now below its 9 EMA.

Crude sold off steadily until early afternoon in the US, where a few buyers showed up that pulled crude up off the floor. After market close, the API report showed a surprisingly bullish inventory draw of -5.8 million barrels, with a gasoline build of +3.1 million barrels. The API report was a mixed bag, but oil ended up moving up about 20 cents on the news. Crude ended the day down -0.30 to 46.32. Candle print was a spinning top, which the code felt was somewhat bullish.

SPX dropped as the result of that end of day selloff, losing -2.46 to 2396.92. It wasn't much of a drop, but SPX still managed to print a relatively highly-rated “bearish engulfing” candle, which the code felt might be a top (57%).   Cyclicals rallied (XLY:+0.58%) while energy led the market lower (XLE:-0.83%) followed by materials (XLB:-0.75%).  It was a bad day for commodity-linked equities.

VIX rose +0.19 to 9.96. A trader friend relayed the news to me that some talking head mentioned that the last time the VIX was below 10 was back in 1993. We are at minimum fear, heaven only knows why.

TLT was down for most of the day, rally in the last 30 minutes to close back almost even, down -0.01%. Candle print was a closing white marubozu/NR7 – a very narrow trading range with a close at the day highs. However, this was not a reversal bar; the candle code felt this print was only slightly bullish.

As TLT rallied at end of day, JNK sold off, closing down -0.08% and printing a closing black marubozu which the code felt was bearish. Hints of risk off from JNK.

CRB fell -0.47%; 3 of 5 sectors fell. CRB remains in a strong downtrend, quite near the lows.

The metals continued to fall today – not just gold and silver, but platinum and palladium too. Arguably, gold and silver did best among the PM metals group. The gold/silver ratio even fell just a bit. Miners are showing some signs of life, highlighted by the GDX:$GOLD ratio which is now starting to move noticeably higher.

That said, we still aren't showing any sort of reversals in the metals themselves, at least not in USD anyway. I'm a fan of buying the dips - and this last three weeks definitely qualifies - but sometimes I'm too early, as I was this time around.  I have to remind myself that it is safer to wait for the reversal bar to show up, because we cannot know for sure how much more selling remains before the turn will occur.

"Oversold gets oversolder, and then gets oversoldest.  And then it drops some more."

Note: If you're reading this and are not yet a member of Peak Prosperity's Gold & Silver Group, please consider joining it now. It's where our active community of precious metals enthusiasts have focused discussions on the developments most likely to impact gold & silver. Simply go here and click the "Join Today" button.


davefairtex's picture
Status: Diamond Member (Offline)
Joined: Sep 3 2008
Posts: 5683
EIA report: bullish; crude +1.45 to 47.33

Petroleum status report showed a bullish crude inventory draw: -5.2M barrels, gas draw: -0.2M barrels, distillates draw: -1.6M barrels - substantially better than the API report from last night.

If we close here, crude will have printed a swing low today.

Like I said, the news in oil may be terrible, but once the managed money longs all get rinsed out - and a bunch more jump in short - I assumed that crude would somehow manage to find a reason to rally.


Giant Elk's picture
Giant Elk
Status: Member (Offline)
Joined: Dec 13 2014
Posts: 20
Candle Code Meanings

Hi Dave,

I have a question if you don't mind. You often quote from your candle code a % chance of marking the low or the high.

The way I would interpret that is the % chance of not taking out the low or high in question from the previous day. But I am not sure how it determines bullishness or bearishness relative to the closing price.

For example, take the plunge in WTIC last Thursday, and the long white candle the following day. Friday's candle I assume would have a very high probability of marking the low, since Crude was up around 5% on Friday and would require a large near term drop to take out the low in question. Therefore the statement "75% chance of marking the low" (I just picked the 75% number for argument) is only relative to a level 5% below the closing price on Friday. It is not necessarily bullish, it just means a large near term drop is unlikely to occur. And you cannot take that as being bullish as at closing price on Friday. (Similarly with the DXY which the code found extremely bullish in today's report).

So would I be right in saying this % value is therefore only useful relative to the magnitude of the previous day's candle? For example if you get a shooting star (no movement on the day) in a downtrend and the code assigns a 60% chance of marking the low, that case would actually be quite bullish since the closing price is pretty much at the low and theoretically there is a 60% chance of price not moving below this level.

Also what is the timeline (a number of days? near term I guess?).

Let me respectfully point out I don't want to read into your code that much, it is your business. I am just curious as I read your report a lot and this one confuses me a bit.

Many Thanks,

davefairtex's picture
Status: Diamond Member (Offline)
Joined: Sep 3 2008
Posts: 5683
candle code

Your interpretation is correct!

Yes the size of the candle does end up mattering a lot.  If you have a very long white candle, I've noticed that its much easier for that to be a low than some sort of shooting star.  Likewise with a very long lower shadow, that's also much easier to be a low too.

Timeline is 10 days.

Yes I agree re: both examples of DXY and WTIC.

All that said, my sense, from using the output to inform my own trading, is that it does seem to reflect a change in momentum fairly well.  My feeling about DXY, for instance, is that once you see a strong swing like this, market tends to move in that direction for a while afterwards. 

What I've noticed is that the two-bar swing calls the code makes are pretty good at picking up both reversals and continuations (i.e. a "swing high" that's also a lower high).  It gets a fair amount of stuff right, so much so that I find myself looking at end of day, and wondering "I wonder what my code will say when I run it".

Single-candle moves do tend to be a little more iffy.

And the low-percentage single-candle calls tend to be the most iffy.  We've seen that in the recent gold downturn.

I've tried changing the training paradigm for the code to show what the expected ROI is over the next 10 & 20 days after a given candle pattern.  I didn't find that output as useful.  It is more helpful to me to know what the trend is, and if its about to reverse.  However, it did reveal exactly what you would expect: the really long white candles don't end up being great buying opportunities; its the smaller candles that are also reversal bars that tend to be the best buys.

Perhaps I should stick more closely to the term "marking a low" rather than saying bullish or bearish.

Sometimes the code says two things at the same time, which can be confusing.  Its certainly confusing for me!  For each candle type (say, a "spinning top"), I have an "up" model and a "down" model, trained on examples where spinning tops ended up being reversal bars.  Each day, the candle code identifies the candle patterns for each item I track.  Then, it asks both up & down models, "what do you think" about the chance of this marking the low.  Usually they agree - on a day when the down model says "2%", the up model will say 35%, or something like that.  Sometimes, however, both models say 40%.  I'm not quite sure how to interpret that!


Giant Elk's picture
Giant Elk
Status: Member (Offline)
Joined: Dec 13 2014
Posts: 20
Candle Code

Thank you for the insights. I take your point that the 2 bar reversal signals are stronger, and a better indicator of future price movement. The 10 day timeline is interesting, more than I guessed so the % is quite significant. I imagine it could be useful in picking a stop loss level if you can anticipate a momentum change. Maybe it works more as a risk management tool rather than an indicator of how much the market may move, ie. focusing on the risk rather than the potential reward.

For example, as you have mentioned recently Silver could produce a sizeable snapback (any day now, any day now, any day now...). The potential upside is quite large given the RSI level and Silver’s behaviour historically. If your code gave a relatively strong rating for marking the low (indicating reduced risk) you’d be looking at an attractive trade with a stop at say 1600 over a 5-10 day period. If the code gave a weak rating (higher risk), well it could still be an attractive trade but that’s where you need to be disciplined with your system I guess.

Anyway, if your model works for you in producing specific calls in specific situations that have significance historically, I think that’s fantastic. I’m not a model guy at all, but I do find it interesting, maybe I could have been in a different life.

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.
Login or Register to post comments