Mon 6/13/11. When will a crash occur?

(Update Wed 6/15/11 4PM EST)
It is so hard to get proper perspective until time has passed an important event. One side of me is extremely pissed that my System has not caught the downtrend of the past couple weeks and that my old bearish bias and System would have largely caught it. But, the other side of me is trying to say that at least the System has not really lost anything either and that there are logical reasons it didn't work in this unusual situation and that since trying to go long on the bounce from 1285, SPX has still only fallen 2% even though the volatility and gaps make it seem much much worse. I have some thoughts about the System's handling of the gap down from 1345 and other key spots as well as the lack of 30% retracements and bottoming indicators, and they largely center around recognition of a rare rapid cycle failure like we had in September 2008, but I think it best to evaluate things in a few weeks.  The cycle is not projected to end until July 19th +/- but there is often a mid-cycle low and 20/50dSMA backtest failure when bearish configurations occur early in a cycle.

For now, yet again at a new SPX low, we have hourly and daily RSI divergences. This time I can't say there was no relief rally in between, since we got a big surge upward on Tuesday. TRIN exceeded 2.0 for the 3rd time in 2 weeks. That happened 9 times in 6 weeks during the May/June 2010 drop but there were 3 violent 1+ week rallies retracing Fib 38%+ of the top during that time span, and we've yet to see that. VIX closed above its uBB20. NYAD is its most oversold in 3 months today with the 20EMA approximating the last several major SPX bottoms. TICK 20EMA has now spent a few days at a level only seen for a few weeks in 10+ years. Stoch failed to crossover bearish today and dRSI14 never fell below 30 with the MACD histogram also somewhat diverging. ISEE=69. That's at an extreme overly bearish low, but to be honest it seems to happen most often near a stock market top or within 1 day of a 1-2%+ multi-day rally. Given the 200dSMA at 1257, key pivot support at 1249 and the bull market uptrend line and other pivots just below that combined with all the bearish sentiment and indicators, it's hard for me to envision the straight down crash scenario from here. I think either SPX bottoms at 1245-1262 and retraces 38%+ of 1371 (currently 38%=1304, 62%=1329) OR it gaps down below 1250 and then backtests 1250-1265 before crashing. Intersetingly, a 38% retrace of 1345 (not 1371) is now 1295. I once started a study to see if price pivots from an important top/bottom could predict the ultimate target by using reverse Fib engineering, because I've seen this happen so many times, but I never followed through. If 1260-1262 sticks, it would be vindication for reverse Fib engineering in this particular case. Overall, my basic point is that under almost all non-crash circumstances and even the vast majority of crash setups, SPX should have a 38-62%+ retrace coming, and now SPX has officially reached my first crash setup criteria with an 8% drop although you'd think it would be worse with 7 consecutive down weeks and such extreme bearish sentiment.

If I were short which I am unfortunately not, I would think large profit-taking would be warranted at 1245-1265 with knowledge that any further big move down at this point will likely lead to a safer backtest entry with stronger risk/reward. I would not be surprised by a big gap up or down Thursday, but the System is neutral in long-short mode with 2-candle resistance currently at 1269. Good luck.

(Update Wed 6/15/11 9:20AM EST)
The large gap down today keeps the daily down trend in tact. Yesterday's rally offered a little indicator relief, but bulls could not bust through 1295. SPX 1276-1279 is the Fib 50-62% retrace level and home of some recent intraday pivots. My expectation for a B-wave pullback to 1282-1287 is dashed. Bulls are hanging on by a thread, and the odds are against them. The System long will likely be stopped out with a small profit and will go short if given a 30-40% retrace. Busy day for me. Good luck.

P.S. I may regret it, but I've decided against selling the long position at 1277-1278 at the open. Instead, I lowered the stop on my long position to the previous gap near 1272. Here's why. When backtesting the new System on the hourly charts, in a situation where an opening gap broke a System s/r level such as happened today, I understood that many gaps are quickly reversed, so I used the extreme of the previous hourly candle that the gap entered. Typically, that allowed a few points of leeway, and, even when the gap day turned into a trend day, the System merely lost a few more points than it could have, but, given so many gap fills, the frequent reward outweighed the losses. In this case, SPX gapped down to the key 1276 area shortly after the open with 1272 not being far below, so I decided to wait and see if the inverse H&S would play out. If SPX were not still early in its cycle having bounced for only 2 partial days after forming a possible mid-cycle low, the System would likely be in short mode and the response would be different. The dollar rally looks promising with Shanghai falling again overnight to its larger triangle bottom, so bulls are certainly against the ropes here at 1277. I'll move the stop up as usual if SPX moves up.

P.P.S. Ouch. Another draw trade (+/- .5%) after a 1275 entry, 25% profit-taking at 1291 and getting stopped out at 1272. The System has been essentially treading water losing about .5% over 4 trades now missing the downtrend. Very frustrating. Original plan this morning to short on a 30% retrace to 1280.7 would have been perfect but not what I backtested on such gaps over s/r. Need to reevaluate the recent weeks for learning lessons. The crash scenario may be in play, but the 26% retrace rally to 1293 in 2 partial days would be tied for both the shortest and smallest pre-crash bounce in the 12 examples I studied after a 7.7% drop which was arguably shy of my 8-10% criteria (actually all 12 examples fit 9-10% rounded so 8-10% was already a stretch), and now we have what looks like a possible 5 waves down from 1371 (may need 1 more low in the coming days if you count the initial drop as a 1-2-1-2-1 instead of an LDT) still in the window for a mid-cycle low. From a pure statistical standpoint looking at the previous crashes I listed, a bottom within the next few days above 1250 with a more sizable bounce into late June or even July 4th would fit the template better, but things rarely go as expected. And, if SPX can make 1266 resistance with VIX holding its breakout, USD rallying and Shanghai falling, there is certainly the setup for a crash. There is an OEW pivot at 1261 +/-. Let's see how SPX behaves through tomorrow's open. Good luck.

(Update Tues 6/14/11 4PM EST)
SPX gapped over the 1279-1284 area that it needed to change the character of the downtrend, and then it pulled back a little from the obvious 1294ish resistance level. Nasdaq, small caps and energy outperformed which supports the rally. Dynamic RSI suggests that the structure from 1266 to 1293 is complete. I suspect a 2nd leg up after a B-wave 20-40% retrace to 1282-1287. The OEW 1303-1313 pivot area contains the 20dSMA and key Fib retracement levels, so that's still my favored target. Above that, there are numerous resistances and OPEX max pain all the way up to 1330 which is the level needed for a System bullish configuration. If SPX can bust through 1295, I expect SPX to chop around the targets above for a few days but ultimately fail due to the configuration of the current cycle and key moving averages as well as all the bearish indicators pointing to a 9%+ top in May/June. The System opened a long position at 1275.38 with current support at 1286.44.

If SPX does rally more this week, 1266 (a 7.7% drop from 1371) becomes the key pivot for a potential crash using the criteria I described earlier. The average crash behavior is to almost always retrace 38%+ (1305ish) and usually 62%+ (1330ish) above the 20dSMA (1310 and falling) for at least 2 days but usually 4-10 days (likely into Friday OPEX or next week's Fed meeting) with the subsequent breakdown at 1265 projected for June 22 - July 8. The System cycle is projected to end July 19th +/-. However, there is a lot of support at 1220-1250, so, assuming this crash setup unfolds, I'd expect SPX to follow other similar historical situations by either (1) gapping through most of 1220-1250 or (2) briefly backtesting the 1266 breakdown after bouncing from 1220-1250. The gap scenario would target 1125-1175. The backtest scenario would target 1100 based on a 1-2-1-2-1-2 count. Don't get me wrong. I'm not predicting a crash. I'm merely saying that the odds are a lot higher for a crash than normal right now, and I'm laying out some of the typical crash criteria, behaviors and targets.

One problem for the bulls is VIX managed to stay above its 20dSMA, but that will likely change if SPX breaks 1295 so that is key. A bigger problem for bears is the extreme bearish sentiment, put/call ratios, oversold NYAD/TICK moving averages etc. I forgot to mention the other day that $TICK 20EMA fell to a level only seen for a few weeks in the last 10+ years and always near a bottom. Although, interestingly, it occurred during the pivot breakdown in the September 2008 crash. It's those extreme bearish readings that support the possibility of either a significant bottom in the coming weeks OR a crash panic below 1220-1250 that has everyone throwing out the kitchen sink. I don't know which one we'll get and I won't blindly bet on a crash, but I'd like to go short on a failed rally. Good luck.

(Update Mon 6/13/11 4PM EST)
VIX briefly pierced its daily upper BB20 and BB50 but has fallen back some. The US Dollar briefly pierced its 50dSMA but has fallen back some. XLF briefly exceeded its Thursday high which is the equivalent of 1295 in SPX, so financials are outperforming. Daily and hourly RSI continued their positive divergences.  However, small caps, energy and tech are still lagging. Since SPX 1345 was reached 9 trading days ago, SPX has just made its 5th 11-12pt rally with one 17pt rally mixed in. Every one of those has failed, so, until SPX makes an 18+ pt rally, the bulls still have their backs against the wall. Currently, 1266+13=1279 and 1266+18=1284 and the lower BB for the daily 50SMA will be around 1280 on Tuesday, so bulls must break 1279-1284 (probably with tech joining in) to make the bears think this rally is different and cover shorts. Until then, SPX keeps churning towards the 1250 level that everybody seems to be watching. I still recommend watching USD and SSEC for clues. Good luck.
_____________________________________________
Before I dig into my crash analysis, let me say that I still have the System in long-short mode expecting an imminent bottom based on daily reversal indicators followed by a sizable dead-cat bounce. Admittedly, a confluence of support looms around 1250 and it's possible SPX could test that first, but it is the exception for the market to continue blowing through System oversold indicators without larger price retraces and/or indicator relief.

When will a crash occur? A crash is highly likely starting somewhere between late June and August. Some people might define a crash as any drop more than 20% or more than 10% within a few weeks, but let me define what I mean by crash. A crash is an 8-10%+ decline from the breakdown of another 8-10% decline's pivot low with a 25%+ retrace in between. By definition, the total decline must be 16%+ but usually exceeds 20% and steepens after the breakdown. To arrive at that definition, I studied the daily chart of every year from 1981 looking for "recognition" moments as part of downtrends in the vicinity of 20%+. A 10% drop, 8% rally and 10% drop would not qualify since the final drop did not continue 8-10%+ below the first drop's low, so that eliminated quite a few possible cases. In terms of "crash" exceptions, I did come across a couple 15%+ drops that did not fit the above crash criteria such as 1440-->1200 in early 2008 when SPX did grind 17% lower on less oversold signals, but they were exceptions and that case did not reach 20%. In the end, I came up with 12 cases that fit my "crash" criteria and one could argue a back-to-back 13th case in summer 2002.


Why do I think such a scenario should be considered and studied now? Well, for months, I've often posted about 10 strong arguments for a May/June top and a likely 9%+ drop into July with further lows or retests into late August or September. Also, pre-QE2 SPX 1011 is such a meaningful target technically and economically, and a drop to such a level would identically match the 2002 and 2008 panics in terms of percent. So, I've been expecting the first criteria for a "crash to be met by July with the second criteria tested (not necessarily triggered) by August. And, there is certainly an economic backdrop supportive of further damage after a 100%+ SPX rally in 2 years. Thus, I decided it was worthwhile to see if I could find any commonality to crashes, so I could know if and when a crash is happening and when to stay short in an oversold downtrend. It is important to mention that SPX has dropped 7.5% from 1371 to 1268 and thus questionably meets the first crash criteria, but it could drop further and you'll see that other borderline cases followed crash characteristics well. Regardless, meeting the first crash criteria has much much less than 50% odds of leading to a crash, but meeting the 2nd criteria does give you something close to 50% odds which is worth trading for the reward.

The 12 "crashes" since 1981 occurred in Aug/Sep 1981, Oct 1987, Aug 1990, Aug 1998, Mar 2001, Aug/Sep 2001, Jun/Jul 2002, Sep/Oct 2002, Jan/Feb 2003, Jan 2008, Sep 2008, Feb/Mar 09. Just by looking at that list, 3 things stand out.
1. The most commonly named crashes are there which validates it to some degree: Oct 87, Sep 01, Sep 08.
2. 58% of the crashes occurred in Aug/Sep/Oct while 33% occurred in Jan/Feb/Mar. Only 1 occurred in April through July.
3. The first 17 years produced 3 crashes. The last 13 years have produced 9 crashes. Can you say secular bear market?

By itself, that information doesn't help us too much except to suggest that a crash, if one is about to occur, is likelier in August. Even a couple borderline cases I found (6-8% drops instead of 8-10%+) fit the same calendar profile: March 1994 and October 2000. But, that is a small sample size and June/July did witness a crash once, so we need more ammo to identify any upcoming summer crash. BTW, the June/July 2002 crash was 26% from its breakdown point (which occurred after an 11% down leg and rally) and that matched the 2008 financial panic 26% crash breakdown and approached the 30% crash breakdown in 1987. Interestingly, 1371-->1015 would be a 26% drop, but a crash breakdown from say a 1260 pivot to 1011 would "only" be a 20% drop and probably not even in the top 5 crashes since 1981.

I was able to identify a few other crash commonalities that could help us after the initial 8-10% drop establishes a pivot low. From the RSI momentum and price low...
(1) Trading day 16-18 was a breakdown day in 8 of 12 cases. Day 7-8 was a breakdown day in 4 of 12 cases as was Day 11-12. Keep in mind the number of trading days from a price low and momentum low to a breakdown day can differ so 12 crash cases can have up to 24 price&momentum lows. There were a couple cases of 24-35 day delays from lows like January 2008 at the recognition of the previous bear market. Also, keep in mind the breakdown day is not the top day. It is the day the key pivot breaks usually leading to a washout or a dip/backtest/washout. So, if a crash breakdown occurs on Day 11 from the previous pivot low, it means that SPX likely topped in its dead-cat bounce after 4-7 days.
(2) No crash occurred without a 2+ day rally first and the only 2-day rally in 2008 was 1133-->1265.
(3) 7 of the 12 cases had 60%+ retracements between the big drops. 3 of the 12 cases retraced about 37-40% (Fib 38%) and 2 cases retraced 33% and 25% with the latter weakest bounce being in 1990.
(4) 10 of the 12 cases reached their 20dSMA between drops. The March 2001 case fell a few points short and August 1990 fell about 2% shy.

In a perfect world where historical stats work today, the average "crash" statistics tell us SPX has its strongest chance to crash in August after establishing a key low at 1220-1260 in July (an 8-10% drop 1-5 weeks before the crash starts) with a 38% or 62% retrace in between. But, there is no perfect world, so a 7.5-8% drop in June followed by a 25-75% retracement and then a June/July crash must be given consideration.

I am going to study the 12 crash scenarios further as well as those that narrowly missed my criteria to determine when a breakdown will fail and when a "crash" trade should be abandoned. Luckily, by my crash definition, I've got at least 7 trading days to do so and possibly 2-5 weeks. I am NOT saying a crash will definitely occur, but we have a high-potential setup because...
(1) My Shanghai Midnight indicator is 70%+ successful and triggered before the 2008 and 2001 crashes.
(2) Fed liquidity support will be largely disappearing at the end of June with SPX 1011 being the pre-QE2 price pivot.
(3) The debt ceiling and Greek debt obligations are within weeks of inflection.
(4) The 200dSMA (last key moving average support), bull market up trend line, previous key pivot lows/highs and the commonly used 10% bear market definition all reside in the 1220-1250 price region setting us up for a vacuum below.
(5) Fear is palpable but with strong belief that a low is near and economic muddle-through is upon us thus setting us up for a panic attack in a moment of "recognition".
(6) A lot of long-term cycles indicate a summer 2011 inflection point.


If 1268 proves to be the low of the down trend from 1345, SPX should rally to at least 1294 and probably 1305+ above the 20dSMA based on previous crash scenarios. If SPX falls lower first which is very plausible, then those rally targets will likewise fall. If a rally does get going, max option pain near 1320 on Friday should be kept in mind since it often serves as a magnet. Also, the Fed has a meeting next week and SPX often creeps higher or consolidates into those meetings. Good luck.