Monday, July 30, 2012

The hard road to consistency III - The heavy cost of education

Trading is very expensive to learn. There are countless ways to trade, many of which are unprofitable and many others profitable only to their sole practitioners. Of these countless ways of trading, only a couple are probably suited for you. Most people who try to trade systems without proving it to themselves will end up losing everything they put into it.

It is my opinion that a couple of observations of market behavior are insufficient to be branded setups and traded off. Its like observing the number thirteen is unlucky or people with the zodiac sign Leo are born leaders. It may seem true due to selection bias but most likely its just simple superstition. Many trading setups are superstitions and you should not trade them until know the probabilities for each setup.

Whatever setup you observe, read on a blog, hear from some hotshot trader or otherwise obtain from other means needs to be tested and measured to determine the following:

  • Potential reward:risk (Needs to be at least 1:1, 2:3 or better recommended)
  • Stop size required (Smaller is better or you may get shaken out)
  • Probability of success: (Theoretically 50%+ but 60%+ recommended)
  • Invalidating conditions and subtle variations (the fewer, the better)

The best way to get this information is to actually trade the setups. A sample of 20 trades will quickly eliminate poor setups. A sample of 40 should be sufficient to calculate probabilities. A sample of 100 should be sufficient to calculate stops.

Setups have subtle variations that make a given setup clear or strong vs muddy or weak. For example, today's trend down b12-27 from the W reversal was broken by a counter-trend move b35-49 and then we got a deep mW b50-55 (two inside bars b52,54 make this three tiny pushes). The subtle difference between the three pushes ending a leg versus an actual reversal of a trend takes experience and many traders could fail to realize that this is not an actual reversal. If they did not take a scalp profit, they probably were stopped out. Such losses add to the overall cost of your trading education and only screen time will make you better.

I believe the first year of trading can be done entirely in SIM and will probably save tens of thousands of dollars for the average trader. I have yet to meet a single trader who became profitable in the first year. Save the capital for once you know what you are doing. Modern SIM trading is extremely realistic and enables you to keep the same detailed records of wins, stops required, etc. The only thing missing is your mental discipline and ability to trade exactly as if it was the real thing. If you wish, you can toss a coin and switch between real trading and SIM trading arbitrarily until your mind does not differentiate.

Most traders fail because they run out of capital and the severe loss of the first few years of trading makes them give up. Reduce the cost of education by carefully evaluating and choosing setups and practicing your system until your prove it to yourself.

Tuesday, July 24, 2012

The hard road to consistency II - The off day

"A Samarian warrior does not fear death, nor does he rush out to meet it"

                                                            --Conan the Barbarian

Trading requires you be at your very best. You need to be alert and focused. You need to be rested and undistracted. You need to have eaten and your body should be healthy. Do not party unless the market is closed the next day. Despite diligence to your alertness and despite years of good trading, you will most certainly have an occasional off day when you are just not in the game.

Many traders who have an early brush with consistency accumulate enough points to think they have made it and start increasing their trading size. This is dangerous because a single off day can wipe you out. The most dangerous thing about nacent consistency is that you won't believe you can be wrong repeatedly on the same day and will have the confidence to think that you can win back everything you lost.

This is why I have a strict two failed trade limit. You may take a PM trade if you lose two trades in the AM but only if a great trade sets up  (for a maximum total of three losses a day).

The following signs should alert you that today is an off day:

  1. You already have a reason to believe you may not be at your best (did not sleep/eat)
  2. You lapsed into poor habits you thought were broken long ago
  3. You are repeatedly reading the market worse than usual
  4. You already hit your max losses

The moment you realize you are off, you need to stop trading. Even many years later, you will certainly encounter a day where you simply are not in tune with the market. Stop before you damage yourself more. Release yourself from the pressure to make money and focus on studying your newfound situation. This day is a poor trading opportunity but a great educational opportunity. Study your behavior and study the market's behavior. 

Friday, July 20, 2012

The hard road to consistency I - Overtrading

The trouble with trading is that you are in a position to do whatever you want and need to struggle to develop the discipline not to do things that hurt you. This is exactly the opposite from the real world where you are in a position to do very few things and have to struggle to do things you want to do.

You will learn very quickly that doing everything that comes to your mind leads quickly to ruin. The only way out is to learn not to do things. Your focus needs to change to stop doing things that cost you money even though they work some of the time.

In trading nothing is guaranteed. No setup is guaranteed to work every single time and even the best setups can have a repeated sequence of many losses. Everything is a probability and probability works only with a large number of samples. Even a 50% setup should be expected to see at least 6 consecutive losers in 100 attempts. (Go ahead, toss a coin 100 times).

Despite having a profitable system, many traders fail and this is because they are unable to follow their own rules. A trader needs to figure out what habits are damaging their trading and root them out.

The most dangerous bad habit is over-trading. What over-trading is depends on your style of trading. If your goal is to trade the 3 to 5 large swings of the day, then any day over 10 trades is probably over-trading. If you are a scalper who wants to take 10 trades a day, 20 may not be that bad but 40 certainly is. The best measure of overtrading is a poor win/trade ratio. 2 points from one trade is decent, 2 points from 20 trades is not.

Overtrading usually is caused by the trader reacting poorly (emotionally) to a loss. There are two kinds of reactions to a lost trade: Retrying and flip-flopping. A trader who has decided the market is going down will continue to short the market over and over, regardless of stop-outs. He thinks he is right about the direction of the market but incorrect about timing or stop size. The next time has to work!

The flip-flopper is far worse. He will go long if stopped out of a short trade and vice-versa. If a great long setup was  stopped out, then the market certainly should drop, right? (Answer: wrong. See b14)

The correct way to react on a loss is to look at the market objectively and enter only if there is an extremely good reason. A trader susceptible to overtrading should just sit out a few bars even if there appears to be a good reason to re-enter. Usually, if you were stopped out of a good setup, the market probably is in chop and you should not trade for a while.

Wait for the price to move away from the area you were stopped out from or give two swings in opposite directions after the bar that stopped you out. Most traders who enter right away after being stopped out have a state of mind thats a possible combination of disappointment, surprise, sense of urgency and desperation all of which are likely to cloud their judgement. Until you are consistent, stay out until price moves away or gives two more swings after a losing trade.

The mind treats losses like a mini-trauma and needs time to recover. Most losses due to over-trading are caused by poor judgement due to recent loss.

You also need to protect your account from being blown up because you cannot accept to have a losing day and will keep trying to trade your way out of a drawdown. Over traders should have a strict rule regarding the number of losing trades they have per day. If you are in a state of mind that  is causing you to lose over and over, trying harder is likely to make it worse, not better.

Understand that your main goal as a trader is to manage risk and once you have crossed a certain number of losing trades, your risk of ending the day with a larger loss is higher if you continue to trade and manage the risk by stopping trading.

Thursday, July 19, 2012

Risks of fBO

An fBO trade requires the highest amount of skill to swing and the least amount of skill to scalp. The reason is that fBOs are expected to give two legs but that's all. There is no expectation of a target. There is no expectation that the other end of the trading range would even be tested.

Since legs can be fairly complex and contain mini-legs, the end of an fBO move is highly subject to a trader's experience and judgement.

For example, the fbo trade below b37 today was expected to give two legs down and it did go sharply to b54. At this point, a trader needs to assess if there would be one more leg to the low of the day or a turn right here.

If the trader correctly guesses the price would turn and buy b54 or the HL at b60, then at b63, the question becomes: will this move end here or was this just a minor pullback in the larger trend down and is likely to result in another move down.

The complexity can be dealt with by assuming that every second push against the traders position forces an exit. For example, b37 sellers would exit above b54. b54 buyers would exit below b63 and so on. Such a strategy requires the trading range to be wide to be profitable. The trading range needs to be roughly the size of a day's move (8-10 points).

An opposing strategy is to trade every fBO as a trend and exit when the price takes out a stop beyond a swing point. This means b54 buyers would hold until their stops below b74 are taken out at b79. None of these are easy and fBO trades certainly need a scalp target since the potential for large swings is undefined.

Tuesday, July 17, 2012

The ideal price action

I have mentioned wide channels with narrow slopes as the most dangerous price action. In contrast, a steep narrow trend with clean bars and no overlap of swings is probably the safest price action and the most profitable.

Clean signal bars (b2) leading to large trend entry bar (b3), with good follow through (b4), shallow pullback that does not overlap prior swing (b6), these are the mark of a very clean and easy day.

On such a day, there are few traps. All signals probably work when strong. The W reversal at b16 or the fL2 at b25 for example, while definitely grounds for suspicion on most days ended up working despite the weak close and overlap.

Until the price action finally became choppy after lunch, the action was clean and clear. If you encounter such a day, take any with-trend signal since it probably will succeed. When possible swing at least one contract to a decent profit.

These days are rare, so make the most of it.

Monday, July 16, 2012

Stop and Scalp size for swing trading

In a previous post, I discussed a way to optimize your stop size. From an extensive history of trades, including various experimental trades, I deduced that my best entries pullback 4t or less after entry. This implied that I could possibly get away with a 5t stop for most trades.
On the left you see a chart displaying the Maximum Adverse Excursion (MAE) of over 100 winning trades. (Losing trades are irrelevant since they always have MAE of where the trade was stopped out.)

The most striking feature of this chart is that most trades are crowded in the first 5 columns (0t, 1t, 2t, 3t and 4t). Entries that have pulled back beyond 4t after triggering and have eventually been profitable less than 10% of the time. Out of those only 3 have managed to give more than 10t profit. I have not made a lot of trades with stops larger than 2 points recently so the sparsity on the far right is probably simply insufficient data. But when I did the same analysis long ago with price action stops (beyond bars) I noticed a similar pattern.

I focus on large market moves and signal bars that lead to large market moves pull back very little after a they trigger. The obviously important turns and their pullbacks today were  b8(4t), b23(2t), b40(0t) and b59(1t). If you notice, they pulled back very little after they triggered. For swing trading, you need to only focus on the large moves and ignore small moves, i.e. your aim is to buy a bar like b40 and hold it till b59 rather than buying the low of b50, b54 and selling at +1 profit each.

Therefore, if you are looking to swing and the price pulls back more than 4t from your entry price, it may not end up being a good swing at all. The next step then is to tune your trading to accurately assess the largest market turns of the day and take only those setups.

Once you have a fixed stop size, then your scalp size is automatically twice the risk if you want to swing half your contracts. (If you want to take a fixed profit, anything larger than your risk is acceptable). This gives a 5t fixed stop and 10t scalp target for my style of trading. Since most of my entries eventually end up being large moves, on an average I'm always better off swinging at least half.

Naturally, 5t/10t above is determined based on my personal trading history and for my strategy of trying to only trade major swings. Your mileage may vary and you would need to determine your stop and scalp points based on your own trading data and strategy.

Friday, July 13, 2012

Breakeven odds

Some of you may have noticed me go from +2 scalps to +2.5. This represents a shift in my risk management strategy that was under extensive testing. Previously, I have posted a risk management system that could be used by traders to take a moderate fixed profit with a small fixed risk.

That system works because every win is larger than every loss and a breakeven trader should still accumulate points. However on some days such as today, the price action obviously favors swinging your position.

Traders who wish to swing should use the following risk management strategy. This strategy is methodology neutral, i.e. it should work with any kind of trading/entry system.

  • Initial Stop is -5t for full size.
  • 1st target is +10t for half size.

Never loosen or tighten the stop until the first target is filled or stop is taken out. If not, you lose the benefit of the math. When first target is filled, move stop on balance to your entry price and let it run. The swing portion can be exited at any price above your entry price but you should try to get at least 4 points.

In the breakeven case:

  • 50% of trades stop you out without filling first target. (10t loss)
  • 50% of trades fill first target and pullback to take your breakeven stop. (10t profit)

This means even if you can never ever swing a trade, the worst you can have is breakeven as long as your win-rate is at least 50%.

If you can manage to keep your win-rate above 50%, your account should slowly accumulate winnings. You can do the following to keep a better than 50% win rate:

  • Trade only after an obvious trend has broken out.
  • Trade only deep pullbacks in the trend.
  • Never take reversals. 
  • Extreme caution between b24 and b60.
  • Avoid poor signal bars (overlapped, tails, too large, wrong color, etc.)

If your win rate is less than 50%, just stop trading anything else except an obvious trend and first get above 50%. A win rate under 50% means you are doing something very very wrong.

Monday, July 9, 2012

Exiting early

Although I prefer to swing most of my trades till a trend terminates, sometimes its prudent to exit early. In general when the market is in a trading range, you cannot expect most trades to really break into a very strong trend and you may need to look to exit early.

A classic application of this rule is an fBO trade. Since fBO can only be expected to give two legs and the two legs may or may not take out any target (i.e. recent swing points or H/LOD), its often prudent to take a large open profit.

For example, today's fBO setup at b31 popped very high for a second leg up at b48. The theoretical swing stop is below the prior swing (b43) and when the price moves close to a potential target (b17H), it makes sense to exit early and re-enter on the expected deep pullback (eventually occurred at b65).  

One option is to exit on strength (inset image) or exit on the next tick against you (eventually below b51).  I prefer to exit on strength when a potential FF type setup shows up (inside bar b49 could act like FF).

Even on trend days, when the market makes a disproportionally strong move, you should expect an extended pullback, which could be either a quick deep pullback or an prolonged shallow pullback. If your swing stop and the price are very far (4+ points) its usually best to take an early profit rather than sit through what could be an prolonged, deep or complex pullback.  

A notable exception to this rule is the late trend, which rarely pulls back.

Friday, July 6, 2012

Channel Types

Channels occur very often and trading them profitably is presently an area of research. In general, there are three kinds of channels and they are traded similarly from a higher level view but the specifics are different.

Trading Range Channel (TRC): A TRC is a shallow sloped wide-range (3+ points between swings) channel made of large bars (2+ points). The move from the open to about b31 today is a classic TRC. For most traders, TRCs are the most dangerous form of price action. This is because they always appear to be just about to reverse and most of these attempts fail. A new trader may be trapped trying to trade today's AM several times: fBO b5, 2L reversal b9, HL after strong reversal b14, W b19. Note that most of these bars were fairly decent and patterns looked clear. a TRC will trick and frustrate traders over and over. These are dangerous to trade because the stops required are large.

Leg Channel (LC): A leg channel has a slightly steeper slope and usually smaller bars. Most bars are however overlapped and there are many bars with tails. The move from b47-59 is a good example. These are fairly easier to trade since the stops required are smaller. You can enter after any pullback or small bar as long as its near the channel trendline. For example, b50, b55 but not b52 since it was away from the trendline. A small poke beyond a prior bar sets up a 1st channel breakout setup (1CBO). The first BO against the channel is a high probability failure so you should take a trade in the direction of its expected failure. You can also fade the close of the first opposing trend bar (b54).

Trend Channel (TC): A trend channel is actually a regular trend move with very small bars. They appear to overlap but each bar dips only a tick or two below the prior bar implying traders entering aggressively on the close of the prior bar. b59-b64 is a very good example. These can be traded like any strong trend. Enter on the first pullback (b63) or the first opposing strong close (b67) or the first tail in the direction of the move (for example if b64 had a 1t bull body).

Channels types are all related to each other. A TRC is a LC on a higher timeframe, which is a TC on a higher timeframe. TCs in turn are regular trends (such as b65-73) on a yet higher timeframe. Therefore trend strength from weakest to strongest is TRC->LC->TC->Trend.

As you see the price action change from one type to another, it may be an advance notice that a trend is likely to breakout as today or when it moves in the opposite sequence, it may indicate that the trend is about to terminate into a trading range.

Monday, July 2, 2012

Spike and inverted channel

A strong spike can give a channel in the same direction of the spike if the price continues to move in the direction of the spike. Occasionally, such as today, we get a channel in the reverse direction of the spike.

A channel in the direction of a strong spike often goes to a measured move of the entire spike. Traders expecting a channel in the bear direction today were probably surprised to encounter the rarer inverted channel. There is one clue that foretells a possible inverted channel: A smaller spike in the opposite direction (b9-11). If the spike is followed by a spike of any reasonable strength in the opposite direction, you would be cautious of a possible inverted channel.

A with-trend channel often (but not always) goes to a measured move of the entire spike. So for a 10 point spike such as today, traders could expect a channel lasting 10 more points down. An inverted channel similarly will reverse the entire spike and is often a good swing at the first pullback (b32) or first deep or 2 legged pullback (b56).