Jan

24

 I had some difficulty finding a concise definition of the DeMark Sequential indicator. From an article written by Mr. Burke in the 1990s and other sources, I constructed a test, but there are variations in calculation and execution, and there are rumors that Mr. DeMark adds proprietary logic.

1) The method A setup for a buy (sell) occurs when there are 9 consecutive bars with closing prices lower (higher) than 4 bars earlier, and the high (low) of either the 8th or 9th bar is higher (lower) than the low (high) price of a bar at least 3 bars earlier in the sequence.

When setup is complete, begin the countdown. Count any bar in which the close is lower (higher) than the close 2 bars earlier. This time, the counted bars do not need to be consecutive.

If the price goes above (drops below) the highest high (lowest low) of the bars in the setup sequence, the setup and countdown are canceled.

If a new setup occurs in the same direction, the countdown resets to zero.

If a new setup occurs in the opposite direction, begin a new countdown and cancel the previous setup and countdown.

If the count of bars that close lower (higher) than 2 bars before reaches 13, and the 13th counted bar closes lower (higher) than the 8th counted bar, it is a buy (sell) signal. It was not clear to me what should happen if the count reached 13, but the 13th bar did not close lower (higher) than the 8th. I decided to cancel the setup and countdown in that case.

2) The test Using the above rules, I tested daily bars of the S&P 500 futures from 1982 to 1989. I checked the net change over various periods following the buy and sell signals. Results of buy signals:

Change next
Date                          12 days 18 days 28 days 42 days 63 days
                    1/3/1984    2.0%    0.3%   -5.2%   -3.8%   -4.8%
                   3/22/1984   -0.9%    0.8%    2.1%   -2.8%   -3.1%
                   7/20/1984    9.5%    9.2%   11.5%   11.1%   12.5%
                  10/20/1987   18.1%   14.5%    7.3%   14.6%   12.2%
                   12/4/1987   11.4%    9.1%   11.2%   11.9%   18.6%

Average                          8.0%    6.8%    5.4%    6.2%    7.1%
Median                           9.5%    9.1%    7.3%   11.1%   12.2%
Average of all periods in sample 0.5%    0.8%    1.3%    1.9%    2.9%

Results of sell signals:

                             Change next
Date                          12 days 18 days 28 days 42 days 63 days
                  10/12/1982   -0.1%    6.8%    1.6%    3.7%    9.9%
                   4/20/1983    4.6%    2.5%    1.2%    6.1%    5.9%
                   6/17/1983   -1.0%   -2.5%   -3.4%   -3.4%   -2.4%
                   7/11/1985   -2.3%   -3.2%   -2.5%   -4.1%   -6.6%
                   3/11/1987    1.9%    4.2%    1.1%   -0.2%    1.9%
                   1/29/1988    1.1%    2.3%    2.3%    0.0%    1.0%
                   3/14/1989   -0.4%    0.4%    4.1%    5.7%    8.4%
                   5/12/1989    1.7%    3.8%    2.2%    4.0%    8.1%

Average                          0.7%    1.8%    0.8%    1.5%    3.3%
Median                           0.5%    2.4%    1.4%    1.8%    3.9%
Average of all periods in sample 0.5%    0.8%    1.3%    1.9%    2.9%

There were only 5 buy signals in 8 years, but they worked out very well, including a buy signal on the day after the 1987 crash. The record of the sell signals was decidedly mixed. The best that can be said is that the 28-day net change was lower than average after a sell signal, although still positive. I decided 28 days was the optimal holding period and considered only 28-day net changes after later signals.

Buy signals since 1990:

 Change next
Date               28 days
     11/15/1991    6.0%
      11/7/1994   -1.2%
      3/19/2001    6.6%
      7/19/2002    8.6%
      5/10/2005    4.2%
      7/14/2006    4.5%
      3/17/2008    9.2%
      1/12/2009  -14.2%
      10/4/2011   13.3%

Average             4.1%
Median              6.0%
Average of
all 28-day periods  0.6%

Sell signals since 1990:

               Change next
Date              28 days
      8/23/1994   -0.9%
      6/19/1995    2.7%
      11/8/1995    3.0%
      9/16/1996    1.9%
      11/4/1996    2.6%
      2/18/1997   -8.3%
      7/31/1997   -4.5%
      3/18/1998   -0.6%
       1/7/1999   -3.1%
      4/23/1999   -4.4%
      11/7/2001    2.2%
       6/3/2003    3.3%
     11/28/2003    6.0%
      1/22/2004    0.6%
     12/29/2004   -1.2%
      9/14/2006    4.2%
       6/1/2007    0.1%
      4/29/2009    8.1%
      9/16/2009    0.3%
      12/8/2010    4.5%
       2/2/2011   -1.5%

Average             0.7%
Median              0.6%
Average of
all 28-day periods  0.6%

The results of buy signals have continued to be on average very good, although also very rare. The results after sell signals appear consistent with randomness.

Anatoly Veltman writes in: 

May I ask, why would 7- or alternatively 8- or alternatively 9- or alternatively 10- or alternatively 11- ….be my guest to go on forever… "work"?

Jordan Low comments: 

I understand where you are going, but your critique will apply to all of technical analysis, and not just DM. I am not a follower of DM, but I believe that technical analysis is based on psychology. At 80F, humans can only live 9 days without water — so there is some cognitive explanation why we are counting 9 days of frustration to capitulate those who traded counter-trend, before the real counter-trend arrives.

Anatoly Veltman writes:

Now, don't ever talk "all T/A". The reason previous volume areas tend to hold the price is because people tend to transact (again) at their former prices. I've always had a beef with time counters who have no accounting for price. The Chair rightfully refers to many as charlatans; but do I understand his page-sized color-coding scheme correctly: the Bond move +0'02 = Bond move +2'00? SP change of 0.50 gets same color as SP change of 15.00??

I bet you that a system that buys a 38%, a 50%, or a 62% retracement of preceding impulsive up-wave will produce better result than a system that buys exactly the "8th declining" 5-min bar, or 15-min bar, or 30-min bar, or 60-min bar or daily bar or weekly bar. Isn't the 8th 5-min bar getting you to where the 4th 10-min bar would get you? What's the magic of counting those bars?
 

Sep

17

Canada, Mexico, S. Korea, USA, UK all have chart patterns that have held the August 11th lows and have zig-zagged sideways to up.

Many more "weaker" countries have broken those lows and are rallying, from lower levels: Italy, Belgium, Switzerland, Malaysia, Netherlands, Austria  , Spain, France, Singapore, Taiwan. I am referring to etfs like ewi, ewk, ewl etc.

So it's west vs the rest. So the question is–are the western lows going to hold and do western markets lead up or do they follow them down into another leg of bear? Is the Euro fix baked into the markets or not? QE seems a given, a Greek default is a given, a two tiered Euro is the only real answer, gold going to new higher highs is a given, then do all markets follow up after a shriek down on crocodile tears bad news (Greek out) ???

I mean everyone knows it's bad bad bad bad….

Jordan Low comments: 

Could Nikkei vs SPX in 1990 be a guide? Both fell until Sept 1990, and rallied from Sept 1990 - Mar 1991. Thereafter, decoupling occurred.

Jul

21

 Tone is as important to music as pitch. You've heard a new violin player playing the notes on the chart but the tone is awful. Yo Yo Ma plays sweet modulating tones. Focusing on just price without regard to tone leaves out relevant and important information.

The market has tone. Quiet, jumpy, weak, anxious, thin, bullish and dense internals. The environment, political, news, economic, social, international also establishes a mood and tone to the market that cannot be ignored except at your own risk. The prior tone is also relevant as music is not discrete tones and notes, but a integrated statement over time. Emotion is the main vehicle of music and also with the markets.

Tone both creates emotion and is the result of emotion. Voice carries emotion: The cry of a baby, the angry tones of politicians, the whine of the complainer each has distinct tonal qualities. The tone carries meaning and has persuasive force. The question is how to quantify it or even qualify it to learn the meaning in the market.

Emotion also has established patterns: Denial, anger, acceptance; fear, capitulation, numbness; catastrophe,catharsis; infatuation, love, boredom, hate. If tone both reveals and creates emotion, understanding the tone of the market will reveal its emotional state and reveal its emotional stage giving a clue to the next phase.

Jordan Low writes: 

The yogis believe in seven chakras in the human body, each corresponding to one of the seven pitches, each corresponding to a different emotion. The stages of grieving you describe go from lower chakras/pitch to higher.

Soros reacts in his gut, which is one of the lower pitches — probably a survival type of emotion. The tone probably is the intensity of the energy at that chakra — I am thinking market volume. We are probably saying the same thing — I find it interesting but am adding quite little.

BTW, the chakras also correspond to the colors in the rainbow, red with the lowest frequency is below.

Laurence Glazier writes:

 For these reasons I find alternating attention between composing and trading congruent, as they are carriers of emotion at different scales.

From my point of view the Elliott Wave framework fits well, being a sequence of eight stages whose final three are labelled A, B, C. As I often use a fractal structure in music, that is another similarity. I may be slow, but it has taken me years to internalise patterns which are lately becoming clearer to see.

Yesterday I put sixty years of the S&P into Advanced GET. The astonishing rise from the doldrums must, in part, be a distortion reflecting the love of printing money, but even were this transcribed from dollars to ounces of gold, I think the ascension of computer followed by internet technology would show how much these developments have added to the wealth of the world.

As I move between Monthly, Weekly and Daily views, the software messes up the precision of my placement of wave counts, and I am thinking to move the whole thing into a graphics program, with the different scales, whether grand or minuettes, callable up via layers. This would help me watch day by day what's going on, a work of art within an art program.

It is impossible to experience in full a piece of music from a short excerpt, and I think likewise the Market, with all its waves and eddies, needs attention from up close and afar.

Also, the rainbow, the universal belief that there are seven colors in it may stem from Isaac Newton's assertion, which was based on his mystic ideas about numerology.

However if you look at a rainbow and count, it is not clear whether blue, indigo and violet are really three colors or two. Also the yellow band is very narrow, though often depicted as equal to the others.

I think context in time is part of what the Market (like a human being) experiences, so as well as volume one might want might look at moving average based indicators, and fractal perspectives.

Laurel Kenner writes: 

Mr. Sogi is exploring an endlessly fascinating topic with his exceptional lucidity and depth of experience. Great performers play the heartstrings by varying tone within phrases. (They also vary dynamics and duration of individual notes in phrases.) They learn how to do this by spending years with master teachers and figuring things out on their own. That's why synthesized music can sound only like an approximation of the "real thing." Because the market is a bazaar of human voices, expressing workaday practicalities, aspirations, fears and strategies, I don't think it's unduly anthropomorphic to look at it as a great performer. And while some of a great performance is spontaneous, much involves muscle memory that training has made reflexive, and must therefore be susceptible of being "sussed out."

Rocky Humbert writes: 

 There's an old game/tv game show called "Name that tune". The gist is that competitors would try to identify the title of a song by hearing only the first X notes; the winner would correctly name the tune in the fewest notes. Human memory being what it is, it was possible to name many popular pieces and classical symphonies by hearing only the first measure of a piece.

However, if one picked a RANDOM measure from somewhere in the middle of the same piece, it was vastly more difficult to identify the title correctly with the same consistency.

This is a reflection of how our memories work; and this phenomenon may have relevance for people looking for patterns in the middle of time series — as opposed to the beginning and ends of time series.

Alan Millhone writes: 

Hello Rocky,

My old friend and top Master checker player, Karl D. Albrecht from Michigan was walking around the playing room full of players at the Tennessee tournament. As Karl walked by many games that were being played into the mid-game he could by sight and memory accurately tell you from what checker opening each board position originated. I found this remarkable.

Regards,

Alan 

May

20

 The Secretary problem, i.e the optimal number of applicants to interview for a secretary job before quitting would seem to have much applicability to the time to come into the markets these days. How many extremes should one wait for in a day before coming in one way or the other, and what is the expectation for such strategies?

Jordan Low writes:

Is it just a coincidence that 1/e is close to 0.38 or the ratio used for Fibonacci time and price projections in technical analysis?

Rocky Humbert writes:

Steve Landsburg recently wrote about a variation on the Secretary Problem. He noted that "an anonymous math department chairman reports on his own strategy for cutting down on the [interview] workload. The math professor believes that one of the most important determinants of a successful career is luck. So each year, the math professor randomly rejects half the applicants without even reading the folders. That way, he eliminates the unlucky ones."

I suspect that there may be a market analogy in this admittedly sarcastic observation.

Phil McDonnell wonders:

Is it a coincidence that Fibonacci believers always seem to use the rhetorical form of: Is it a coincidence that (fill in a single observation) came close to (fill in selected Fibonacci value)?

Russ Herrold replies:

I do not think there is a co-incidence– the human mind tries to order data, and Fib sequences crop up everywhere. It is natural to do a 'trial fit' just as the eye tries to estimate a fit for a curve [and thus the reason for ready transforms as normalization, log, 1/exp and the like in running a regression– scaling often permits identifying the 'noise' and getting a 'good enough' solution]

Doing the math, of a run of repeated application of any two integers (seemingly separated by whatever distance, although I have not done a formal analysis or proof), the series seem to converge to a Fib set reasonably monotonically after say five rounds for low integers.

I ran into Fib numbers, learning the run time pass estimates for the IBM sort-merge algorithms in in the late '60s, and it appears that Knuth found them in sequential pass sorting as well. I seem to recall a childhood cartoon called 'Donald Duck in Mathmagic Land' where that quackish fellow pointed out that golden spiral, and the perfect rectangle 

Apr

6

 One has always hypothesized that holidays are inordinately associated with major turning points. One hypothesizes that the correlation between the extent of bailout and subsequent economic recovery between countries is not zero. One notes the story from The Book of 5 Rings where a group of wealthy samurai were traveling in Kyoto and were met by a vanguard of vassals telling them that a group of noblemen were behind them and they should bow down on floor in prostration. It turned out the noblemen were robbers and stripped them of their clothes and honor and the samurai had no recourse but to renounce their profession out of shame. What lessons does this have for markets, market people, and others? One believes that the early leads in basketball games and other games tend to be increased in subsequent parts of the game. One hypothesizes that the expected change from the time that the NBER announces that a recession is over or started are opposite in direction from the economy's current announced state — i.e., after they announce a recession the market goes up more than after they announce an expansion.

Rocky Humbert writes:

The continuous surfeit of negativity over the past year (and now hints of protectionism) makes one ponder whether one fell asleep during the housing bubble and awoke in Bizarro World (the mirror-image of Superman's world)… A successful investor doesn't need to either celebrate unemployment (Dr. Rehmke) nor declare millions are out of work forever (Dr. Dreyfus). Both statements are simply provocative– it's much less dramatic to simply observe that employment is a lagging indicator. (Yawn.) Perhaps it would be good medicine for all– if the Chair resumed his slights toward Alan Abelson (last mention July 20, 2007), and prior to that more than once/month. Most importantly, for those who are looking for a "major turning point," I share the words of Bruce Kovner, with whom I had the honor of briefly working: "Listen to the market."

Put simply: for the past year, the optimists saw a monetary/fiscal cyclical recovery with the yield curve predicting growth and inflation. During the same time, the pessimists saw a false stimulus/inventory uptick with the yield curve predicting troubling deficit/supply overhang. (No one expects the job market to recover meaningfully before 2011+.)

By any standard, the pessimists have been horribly wrong. But instead of acknowledging that things are improving, they are being Alan Abelsons, digging in their heels, and predicting that the next huge downturn is just around the corner.

Pitt T. Maner III comments:

Trying to be more in the optimistic camp it looks like the US unemployment fits a hysteresis model. Sort of like TW at Augusta and at home–it could take awhile and that's if there are no more more shocks along the way! :

1. According to Caporale and Alana , two well-known facts about the unemployment rate are

(i) the high persistence of shocks, or hysteresis (see Blanchard and Summers, 1987), which is a feature, among others, of "insider" models (see Lindbeck and Snower, 1988), or of models in which fixed and sunk costs make current unemployment a function of past labour demand (see Cross, 1994, 1995), and

(ii) its asymmetric behaviour, namely the fact that unemployment appears to rise faster in recessions than it falls during recoveries.

2. The next survey of Professional Forecasters will be May 14th, but most see an improving jobs situation. Slow at first but accelerating by end of current POTUS term.

3. Interesting chart of GDP fall vs. unemployment rise (Okun ratio). Less regulations in US (and Canada) would seem to be a long term positive, but the US and Canada sensitivities to GDP fall are higher because workers can be let go faster.

Kim Zussman adds:

Here is an update on P/E type-forecast, with a caveat about markets remaining irrational…if they feel like it.

Pitt T. Maner III comments:

What about the case where you may be moving quickly from high P/Es to lower estimated 12-month forward P/Es? (i.e. S&P 500 going from 31 one year ago to 23 now to possible 15 in 12 months time). So if you have a high rate of change in the P/E downward (if numerators continue to grow) that might make the positive portion of the bars more likely?

It seems with the P/E in the 15-17 range you have more variability in range of returns but the forward dividend yield would still indicate lower returns given yield of 3.4 (one year ago) to about 1.8 today.

Jordan Low replies:

I think that the 10 year data mines the worst case as it includes both the dotcom and subprime busts. The peak of the dotcom was almost exactly 10 years ago, and investors in 2000 weren't looking for E. (They preferred g.)

I understand that the long window is supposed to average the business cycle. Well then, the window should be variable. As of right now, we are getting bad earnings from two crashes and many of those companies don't exist anymore. Perhaps it says something about the unfortunate timing of two bad periods and growth of the Internet being captured by late comers (e.g. Google and Facebook) rather than early adopters (AOL and Yahoo). Being an investor in today's market, not yesterday's, may still be attractive.

Mar

7

Japanese geishaSP500 weekly returns 1/84-present were checked for mean and stdev:

mean 0.001671
stdev 0.023238

Random number generator was used to generate 100 simulated 36-year markets, with the same stdev (0.023238) but with a mean weekly return of zero. The means of all 100 simulated markets were ranked, and the highest one was found to be 0.001663.

So the probability that actual SP500 weekly returns averaged what it did - 0.001671 - by chance alone - was <1%.

Over the same period for Japan's Nikkei 225, the actual mean and stdev for weekly returns was:

mean 0.000455
stdev 0.029022

Using the same method as for SP500 simulation, the random number generator was used to generate 100 simulated 36-year markets, with the same stdev (0.029022) but again a mean weekly return of zero. The means of all 100 simulated markets were ranked, and the highest one was found to be 0.00174. The actual mean of 0.000455 ranked 30th out of 100 simulated weeks. Japanese stock market drift had a 30% probability of occurring by chance alone.

If upward drift was the result of return for risk, why didn't it occur in Japan?

A friend suggested evaluating drift after adjusting for risk-free return (in this case, 30 day t-bill rates available as concurrent alternative to SP500 index investing). Weekly 30-day t-bill yield data from FOMC* (which uses annualized yield) was converted to weekly yield, and SP500 weekly returns were converted to "risk free return" by subtracting the weekly 30 day yield.

SP500 mean weekly returns (1954-present), and SP500 weekly "risk free" were compared to zero with t-test:

One-Sample T: wk ret, "risk free"

Test of mu = 0 vs not = 0

Variable        N         Mean      StDev     SE Mean          95% CI            T          P
wk ret       2929  0.00152  0.0210  0.0003  ( 0.0007, 0.0022)  3.93  0.000
"risk free"  2930  0.00055  0.0210  0.0003  (-0.0002, 0.0013)  1.43  0.152

Subtracting risk free rate of return dropped the return for SP500 by about 2/3, after which drift is no longer significantly different than
zero.

Jordan Low comments:

Are all dividends included? Perhaps it is because Japanese hold stocks for different reasons? I was in Tokyo when JAL went under and some people were happy to pay 1 yen to be a shareholder to get a certain number of tickets for half price a year.

Feb

11

Mr. Galen Cawley kindly provided a link to a dated, but still interesting article entitled, "An analysis of the profiles and motivations of habitual commodity speculators". This comment was on William Weaver's post: "Study shows why it is so scary to lose money". I found this article provocative because when I was younger, some of my innate instincts were similar to the habitually losing commodity speculators'. I suspect everyone who reads this article will see some foibles he has overcome (or needs to overcome) for self-improvement. To summarize the article:

1) Most speculators use too much leverage.

2) Most speculators don't hold their positions for sufficient time.

3) Most speculators don't use stop loss orders. (However, the authors didn't differentiate between actual stop-loss orders and mental stops. The point was they don't cut their losses.)

4) Most speculators prefer a serious of "modest" short-term profits versus slowly accumulating long-term gains in a single position.

Bottom line: They found that the average speculator had a win/loss percentage of 51.3% and the best speculator in the study had a win/loss percentage of 80%. However, he still lost money because of his low profit factor. Despite great win/loss ratios, the average trader in the study is a career net loser. It's a Wall Street platitude that "no one ever went bankrupt taking a profit." This study shows that the platitude is false.

Kim Zussman comments:

"Most speculators don't hold their positions for sufficient time."

"Most speculators don't use stop loss orders… The point was they don't cut their losses."

Which is contradictory because many deep losses reverse eventually ("every price is hit twice")*, and often the correct trade is to wait for losses to reverse (and often it is not). Taken together, this means "ride your winners and ditch your losers." Isn't that trendfollowing, and if so, what if the (currently traded) market is not trendy?

*except Nasdaq 5000 in our lifetimes.

Victor Niederhoffer comments:

One has not read the article, but would wonder whether speculators would lose as much if they showed opposite traits. However, the characteristics noted all would lead to greater vigorish and this is where most of profits from the market makers occur. Everything that happens is guaranteed to increase the profitability of that vigorish to the house or top feeders.

Rocky Humbert responds:

If one attributes the speculator's losses entirely to transaction costs, then one's performance can be improved by simply having fewer transactions. One wonders whether there is an a priori relationship between all of w/l ratio, profit factor and quantity of transactions? Splitting hairs, I quarrel with your choice of the word "guaranteed." A much wiser man taught me that the only things guaranteed in life are "death and taxes."

Jordan Low comments:

Macro traders should be long gamma as markets can stay irrational, while value traders should be short gamma especially if they are positive carry and see no devaluation/fraud risks. I think it all depends on your style of trading.

Victor Niederhoffer writes:

One would say this is perfect as one trader should always be one way and the other trader should always be the other way. that is guaranteed to make the top feeders a perfect guaranteed profit. I say the above in all seriousness and respect to the high thinking Mr. Low.

Rocky Humbert comments:

Mr. Low articulately differentiates between arbitrage/positive carry, mean-reversion (providing liqudity) and trend-following (taking liquidity). Arguably every single trade or investment fits into one of these categories. But I'm confused and intrigued by Vic's comment about "top feeding".

Specifically:

1. Who exactly are the "top feeders" and should we buy their stock?

2. Are the top feeders always the same participants? If they always make a "perfect guaranteed profit" then how can they ever slip from their perch of being a top feeder? (Were EF Hutton, SW Strauss, Shearson, SmithBarney, Dillon, Kidder, Drexel, Revco, Bear, Lehman, Merrill once top-feeders?)

3. And if they do eventually slip off their perch of "top feeder" — then how does one argue that they ever had any advantage? As opposed to being just another opportunistic profit-seeking participant?

Theoretical questions of course — but they challenge the hypothesis.

Yishen Kuik comments:

Seems to me there is some kind of law of conservation going on. Usually if win-loss is very favorable, avg win > avg loss, then frequency of trades is low. If win-loss is very favorable, avg win < avg loss, then frequency of trades is moderate. If win-loss is only very slightly favorable, avg win > avg loss, then frequency of trades is high.

etc. etc. …. all solving for a reasonable profit factor.

But if you are doing something quite different or if you are early in a wave, you can have an unreasonably high profit factor. Then again, there is always the argument that low profit factor strategies with low capacity are the ones with the longest half lives.

Jordan  Low replies:

I appreciate the response.

My view is that the trades do not always net out. While the macro traders are attracted to events such as subprime, china, commodities or Greece, value traders avoid those trades rather than taking the opposite view. Buffet did not invest in dotcoms for example. Each has his own reasons, from only trading what he understands, to seeking or avoiding volume/volatility, to seeking catalysts.

Aside from trades, there might also be a timing mismatch. Traders with most skill in my view know when to sit on their hands. Others might be lucky or unlucky. Eg penny stock momentum strategies might only work in the dotcom era but not after. On the other hand, value did well in the great moderation of 2004-2006 but not during the quant meltdown of 2007. Given the tendancy to chase performance, the dollars in each camp may not net out exactly.

 I hope this clarifies my view, but I do think there are top feeders that make good returns off their franchise. ETF providers make money both by providing access to hot markets and by lending stock to short, for example.

Call me Jordan…

Feb

10

One theory is option pinning as we get to expiration. Another theory is to maximize trading volume as emotions rise on either side of the round number. In a zero sum game, the smart money make more if they increase the magnitude of the other summand.

Archives

Resources & Links

Search