atto

I don't think so, at least directly. Black Swans may be caused by small or big things. The Butterfly Effect, from what I understand, seems to work the best theoretically, because it's impossible to identify the cause that started the effect (if the cause is indeed extremely tiny). "Black Swans" aren't effects, theories, or laws. There is no theory guiding or modeling them, besides "expect the unexpected". On another trading forum, I saw someone had identified three Black Swan potentials: S&P at 1000, Oil at $200, and a big financial going bankrupt. Sounds great, except those aren't Black Swans (and when I read the thread, not one person had noticed this). Black Swans aren't just unlikely events that we may have some forsight; they are extremely sudden, extremely important, and extremely unpredictable. With the real estate bust, credit crisis, bear markets, etc, it wouldn't be "shock and awe" for any of those identified "Black Swans" to happen.

Unfortunately, I don't know of any great educational resources of uses of Mathematics packages for financial analysis. To give you an example of what you can do, I have been working on my gap trade method for a while. The age-old gap trade involves you fading the gap, closing at the last market close, or closing the position at market end for a loss. This has unlimited risk, finite gain, and trades during unfavorable times (gap-and-go's, where the market opens and kills the fade, happen every so often). Basically, gap trades can be "easily" improved by adding some rules, such as implementing risk control (stop loss), and finding biases when gaps close a) more often, and b) for a higher p/l. So, I take a bunch of market data (this one can be easily done with 5-min data) and try to find biases. Do gaps close more often on certain days of the week? Do certain gap sizes affect if it closes? What is the optimal risk/reward? What methods of risk control should be used to keep as many winners in and losers out? Does gap size in relation to yesterday's range have an exploitable bias? Does the open price in relation to value areas of yesterday help? etc, etc, etc I have no formal quant training, so I stumble around until I figure out what works. I recommend for you to read Way of the Turtle, especially the chapters on avoiding curve fitting. You're trying to find biases, not curve fit the thing to past data hoping it continues to work. For instance, gaps bigger than 10 ES don't have enough data to analyze, so I do not trade them (even if some were exploitable). The variables I optimize peak at certain points (such as stop loss size), which makes me confident I'm not benefiting from a few trades but an overall trend. I've been working on my gap trade for a while, and I plan on writing it up (along with the methods I used to find them, including any source code), when I am comfortable with it.

I was able to catch this too shorting the futures. For these kinds of trades, I use my "high volitility - higher risk" risk rules; lower position size, and aggressive chasing stops. Luckily, I got in at 1334 and had absolutely no heat. I got out at 1324 nearing support from earlier in the day and as intense selling pressure started to die. Managing risk and the trade is important for these kinds of trades, as you mentioned, because when they go against you, it can be brutal.

Yeah, this is probably doing the same thing. Basically, we've identified that volume acts in certain ways at micro-tops and micro-bottoms in relation to the number of orders. On the last chart I annotated, a short/long entry could easily be improved by a few ticks if you pay attention to this. I'm not sure if Delta has any advantages, because it doesn't nessessarily tell you anything else (though it may help with a direction bias). What if we were able to identify these "wall points" on a time chart with a symbol on the bar, like an asterisk? In addition to POC, this could give us a very good idea of value and market flow. Thoughts? (As a note, something I want to stay away from is busying up a chart, without adding much additional value.)

Gotcha, I fully understand what you're saying. I use the same technique for entries / exits often. Something you hit on one of your charts is using volume on a tick chart, which in effect, shows order size for the tick bar. I highlighted in your attachment significant volume bars that peak above the rest for the benefit of everyone else. Notice what happens to price action after a volume peak. Almost every micro-wave was identified extremely close to the high/low (within 1-2 bars and 1-2 ticks on the chart). I've used this for entries / exits for a while, and I think it's a very valuable tool on the micro scale (I wouldn't trade based only on these, but they help get a good price). Any ideas how we could integrate this into a standard time chart?

Besides the uploading image bug (which is known about), I have noticed another small bug on the book reviews. The replies suffer from an off-by-one bug, where the first reply is not labeled, and the second is labeled as Post #1. Besides the different numbering, the first reply does not have a perma-link to it (because it has an index of 0, I'm assuming). The bug is active on my book review thread here.

I can see that. He is arrogant and makes attacks on the French several time . Outside of that, he's a smart fellow. I'll check out Rebonato.

The flaw with candlesticks is that they weigh the open and close, when intraday, the bar open/close means very little specifically. What I mean by that is depending on what time / timeframe is governing the candles, you could end up with different stories regarding price. I could create an artificial example, but I'm sure you know what I'm saying. If in your tick chart you had 10 extra ticks in the beginning, those wicks could easily be closes. Furthermore, bodies you see could be wicks depending on when the bar closed. I'd love to see a way to display data that gives weight to total value for that bar (if it's far away from a wick, you know price was rejected, verses happening to close down and exposing a wick) in a friendly manner. With this said, I think there is some value in candles (and the ability to ignore wicks) for daily and larger timeframes, as it represents a market open to close.

You're absolutely right; I didn't think of that. I'd love to assist with development if you want (or even just testing). I use Ninja as well. I like that POC candlestick idea, and I have a few others to add to it. Let me think about this, and I think we can come up with some innovative ideas.

What is your goal? Are you trying to win, do better than half, or what? Your strategy may change depending on that.

Perfectly valid gripe about the book. However, I'm not sure he would tell a trader to ever go with their gut. Since we're dealing with patterns, we can very easily be duped by our brians into seeing what we want to see. I wouldn't parallel the book directly to trading, but use it as a constant lesson that "less is more". I'm not sure the exact primise is applicable to trading.

Risk experts do not understand risk. The Bell Curve has been abused since its inception. You are exposing yourself to risks you don't even know about. Nassim Taleb's hit followup to Fooled by Randomness: The Hidden Role of Chance in the Markets and in Life exposes the academic and risk community for failing at modeling risk, while denying he's any better. Taleb focuses primarily on theory and ideas rather than direct trading application; however, this book is extremely valuable to anyone who deals with risk (read: you!). This book is written with the non-technical in mind. Black Swan makes the argument that the outliers that end up shaping much of our finances and lives are unpredictable and unpreventable. Distributions are tending towards the extreme and irregular verses limited and regular. For example, the stock market crash in October 1987 was something that should have been statistically impossible, but it happened. Russia's default on their bonds in 1997 should have been a far off possibility, but it happened. Taleb argues that these "black swan" events have a termendous effect on us, but it is foolish to even try to predict them. Just because there are a million white swans, it doesn't mean all swans are white. Ergo, just because it's never happened, doesn't mean it never will. The book itself is filled with humorous anecdotes, and often goes off on tangents. Taleb makes very clear his distain for academics, economists, and those who calculate risk. At times, he comes off with an elitist and arrogant air. However, the points he makes are refreshing and insightful. Taleb encourages readers to try limit their downside exposure in all things (not just trading), even against the unlikely, and leave their upside exposure open. That way, readers stand to benefit from black swan events. All in all, this book has reshaped my view of risk. I think The Black Swan is highly insightful and a great read. Check it out on Amazon.com.

Okay, I understand what you're getting at. You're considering an upward-biased market (according to your own analysis) a different good than a downward-biased market (same). Under that line of thinking, I agree. There are biases that cause me to value the market higher than it is, so I trade. Thanks for explaining that out.

Could you elaborate on your comment regarding the "good" (of a financial market) changing? I have a very firm grasp of economics, and I've never heard anyone say that, so I'm open to your theory. I've always believed that the "good" in an ES contract is simply $50 x S&P 500. The S&P 500 is a "good" comprised of the biggest market players, which are goods representing companies. That "good" doesn't change. The value of it may change (such as when Steve Jobs joined AAPL), but that is then reflected in the price. In your example, how does a confirm double bottom change what the ES is? It seems like it's exactly the same thing, just at a different valuation. That is a function of supply and demand, which would also govern the price of your basketball.

charcoalstick, the problem you'll deal with is that not all trades are created equal. You have to discover what works for you. For instance, if you're a trend trader, a good filter might be 5 period MA above 30 period MA. This, of course, wouldn't be used alone, but would do a good job keeping you out of bear markets (for longs). You may look for overbought/oversold signals if you're a reversal trader. You will be pressed to find a system that just uses conventional indicators (RSI, MACD, MA's, Stoch's, etc) and is profitable. At the end, you're the person pushing the button. Therefore, screen time is extremely valuable. Speaking of indicators, try not to fall into a quest for a holy grail indicator. To my knowledge (and trust me, I've looked), it doesn't exist. If you create/discover one, I advise you to keep it secret . Personally, I don't use indicators. Some people do. Again, try to find what works for you.

I used to be very politically minded and get engaged in political discussion at large. I have since realized that it is near impossible to get a truely honest and brilliant man in the oval office. Instead of benefiting from solid ideas and reform, politicans mostly pander for votes. For instance, I've yet to hear what the two viable canidates plan to do about government spending or Social Security. I used to "support the war". After educating myself a little more, I think it's sick, and the unintended consequences will be greater than anyone could have imagined. However, my "support" for the war has nothing to do with my love for America or support for the troops. To throw those uncorrelated and unconnected ideas together is just stupid. In my opinion, we simply do not and cannot understand the politics and culture there. It seems George Washington was right:

This is a wonderful point. I recommend anyone who takes any risk (everyone) to read The Black Swan. It's truly insightful and thought provoking concerning risk and how bad we are at managing it. (P.S. I absolutely love it when people use the alternate definition of "affect" correctly, assuming you meant to .)

Interesting. I bought Mathematica (at a super discount, so I don't feed stuck to it), and it has SQL connection ability (so does Matlab). I could import all the data to a mySQL database, and then use that. mySQL is very efficient at retrieving data, so this may be the way to go.

Wonderful information. What data source do you guys use in any of these mathematical packages? I'd really like to get a few years of ES data (as small timeframe as possible, but 5 minute would be great) on a continious price adjusted contract (price and volume is a necessity). The main bottleneck I've noticed with Excel is working with large data. It handles small data just fine, but I was doing an analysis on just 5000 points (abet, slightly complicated analysis), and the chart took a while to draw and would lock up occasionally.

I've put much thought into doing this as well, but I realized that no matter how you display the data, you're never going to find "new" data from it. The closest exception to this I can think of is Market Profile, as it changes how volume is displayed over both axis. Outside of that, you have two main pieces of data: price and volume. You can distort, contort, and retort them however you want, but that's all we really have. For example, range bars "simplify" price action and enable us to identify breakouts and trends easier; however, the exact same information is just as available from a normal time chart (the high's and low's are the same, as is the general price action). Volume charts give us a good indication of price over volume; however, all that is changed is relative timeframe (as close price action is occasionally "closer" or "further" out due to volume), and the data is the same. Visually, some charts may help you identify what you're looking for, but you're looking at the same side of the same coin. Almost anything you deduce from range bar, volume, tick, time, PnF, etc charts could be deduced from another (given no data smoothing or loss). Market Profile (and any other chart with Volume at Price), as I mentioned, is different because it opens the y-axis for the display of volume. This is novel, because it allows us to visualize volume in a new way. In my opinion, you'd need an innovation like this to significantly improve a chart.

In my opinion, the YM is great for beginners or anyone trading smaller sizes (or is able to effectively scale in and out as to not disrupt the market). Fills are wonderful, and the price action is nearly identical to the ES. Markets also have "personalities". The ER2, for instance, jumps around and moves freely, and once it breaks out, it's gone. The NQ seems to be the most directional of the US index futures. The ES is like a giant train. It moves slower, and will often come back before breaking out. The YM is a moderate choice, which can make it easier for starting traders (of course, this is just my opinion). The ES will get you, as I mentioned above, with its fills. Since ES ticks are 2.5 times the relative size of YM ticks (0.25 pts ES is roughly 2.5 pts YM), price action is a little more "blocky". That extra size hurts fills.

Ah yeah, MATLAB; I've used that before. I was looking into something that looks similar.. Derivates Expert for Mathematica. I know I can get a discount on Mathematica, which is why I was attracted to that (no idea about MATLAB).

Yeah, if your goal is to win, you'll need to take some big risks (even some with slightly negative expectation) to put you ahead. These kinds of trading competitions promote very little "good" trading/investing. If options are available to you, use them (you'll be able to get paid for more risk prema). I won an economics "investment competition" once by lobbying for use of derivates ("But, real investors and traders can use them!"), and then loaded up on risk. Classmate's account ballances were not known, so I had to guess how much I'd need to win. I ended up winning by 140ish% after some very lucky SPY OTM options. The whole thing was a joke, because grades for the assignment were handed out on a bell curve based on account performance (top 20% got A's, next 30% got B's, next 40% got C's, and the rest got D's; if you didn't trade, you failed ). Additionally, The winner got 5 bonus points on the next test. My overall strategy was to shoot for first, settle for a B (the top 50% got A's and B's, and it's extremely simple to make that group).

If I may ask, what programs are you using to do your research? I'm considering buying Mathematica to do some larger data analysis, and don't know if there's anything better on the market currently. What data feed are you using? I heard Excel can do wonders, but I come from a programming background, so often I find it easier to write a program than sludge through menus.

Correct. At the end of the day, 1 YM is about the same as 1 ES when you take into account $/tick and range. However, the ES has significantly more volume than the YM, so if you are trading large size (or off-hours), the ES can be much more favorable. My main complaint with the ES is that fills aren't great due to the larger relative tick size (1 tick ES has the approximate effective range as 2.5 pts YM). I would really like the ES to switch to $0.10 ticks, but I doubt that'll happen.