Caution: Math Geek at Work

A bit more explanation on the reasoning behind the trade I'm in from yesterday's post. The trade is based on two results from my last research project (go to my site, click on "Research Archives" and it's the project at the top dated January 2008).

1. Trend extension legs like the HHLL we've just had on EUR/JPY are followed by legs with a median pullback of 60%. Such a pullback would be a LLLH which is generally followed by a trend extension (down in this case) of 110%.

2. In my notes near the end of the first section, there is a part where I say "Note the weirdness." This is in reference to what looks like a very non-random behavior by which uptrends and downtrends tend to alternate (instead of continuing) after interruptions AND the type of interruptions to trends (range expansion or range contraction legs) also alternate. We have seen a downtrend interrupted by a range contraction which was then followed by un uptrend interrupted by a range expansion, so this follows that pattern. If the pattern continues, I expect that we'll next see a series of downtrend legs followed by a range contraction.

Downtrend legs are shown in orange, uptrend legs in cyan, and interruptions (range expansions or range contractions) in white. Anyway, that's why I'm bearish here, even though we're in a short term up move.

Here's the updated situation on the trade. The 4,5,6 and 7 unit sell orders have filled as the price moves north. The trade is well within expected parameters, currently at a 100 pip loss on a total of 4+5+6+7 = 22 units. Below is the zoomed in annotated view of the action.

I won't keep posting charts on the same trade but here's an update. The move up today in EUR/JPY led to a fill of my 8-unit order, bringing me to a total short position of 30 units now. Price is about 70 pips above my new average entry price as I write this.

Down 0.7% this week due mainly to the open trade on EUR/JPY which is still in negative territory. More on that trade in the next post. Here's the account equity.

The EUR/JPY continues to move upwards, causing one more of my limit orders to fill for 9 more units short. This brings the total position to 39 units which is 1/2 of the maximum position that can be created from this trade. Price just bounced off of the lower of two resistance levels and this upleg is well past the median length of 60% of the previous HHLL leg.

Will resistance hold? Or will price blast through it and smoosh me?

Well the resistance didn't hold, so both of those purple lines are now gone. Two more orders have filled for 10 and 11 units, bringing my total short position up to 60 units.

Most significantly though, the situation has now changed. I set up the original trade after the completion of the last range expansion leg which had gone from a higher high to a lower low (HHLL). The most likely (p=82%)outcome for the upward leg that followed was a failure to penetrate the prior high. However, it just DID penetrate that high, forming a LLHH leg or in other words another range expansion leg. Two RE legs like that back to back aren't common.

Crap, so I may very well be short at a major bottom here. The median pullback length for a LLHH leg is about 60% which gives me a 50% chance that price will retrace to 156.30 if the high was 159.60. That would still put me in profit, and at this point my focus is going to be on getting my stop to B/E as soon as possible and getting out of the short.

Price has reached to within 90 pips of my stop though, so the probability of a loss on this trade is rising.

Cap >>> <<< EUR/JPY

On the brighter side, I had my seventh straight winning scalping session yesterday. Three more and I'll have earned the right to trade larger. I'm loosening up the rules for this a little bit, because it would have really sucked to have gotten nine winning sessions in a row, and then a losing one which puts me back at the starting point.

So the new rules for increasing size are:

1. Have 10 winning sessions without exceeding 3 losing sessions.

2. The winning sessions must have a cumulative total of at least 100 pips

3. The losing sessions must have a total of no more than 30 pips.

4. If I exceed three losing sessions without reaching 10 winning ones first, then I have to start the level over.

5. If I get three losing sessions in a row, then I have to drop down to the next lower base unit size.

6. If increasing unit size results in risks that are greater than 2% of the account size, then I cannot increase size no matter how well I'm doing.

The number of pips in a session is a function of my "base unit" which is the thing that I'm trying to increase in size. But no matter the current size of my base unit, I can trade more than one in any given session as a result of scaling in. So for example if I go long just one base unit, and get 7 pips on it, then that's a 7 pip gain. But if I end up long 3 units and get 4 pips on that trade, then it's really a 12 pip gain based on my "base unit."

Here's the little table I set up to track my progress. There are 10 levels of position sizing shown here, although once I get to the tenth there's no reason I couldn't just keep going. The "BU Factor" shows the base unit size at each level as a multiple of my lowest size which I'm trading now. So if I get my 10 winning sessions, I'll double my position size, then move it up to 5 times the original, then 10, then 20, etc. But I have to get through each level first.

Since I'm in job-hunting mode, I'm not scalping every day. This was my first session in over a week I think. Hopefully I can get on a regular schedule soon.

The account is up 3.2% this week due to the strength in the kiwi. I have a long term low leverage long position there. My open short on EUR/JPY is still in the red though.

Stopped out of that EUR/JPY short for a loss of 1.7% of account value. Bleh. So much for that experiment. Next I'll look at placing stop entry orders behind the price as it moves along a leg. That way I don't enter until it turns and starts to head in my direction.

On the scalping front I now have eight profitable sessions and one session where I broke even. I put that one in the loss column since there weren't any there yet (it could go in either one). So now I have to get two more winning sessions with a total gain of 7 pips in order to complete this level. Then I've earned the right to increase size.

Even though the scalping looks successful, the sessions are kind of stressful so I'm looking at creating a mechanical method for this where I can just set and forget the orders. This has its own problems though because when I'm watching the screen I can use a wide "catastrophe stop" and decide to exit at a loss well before that point. So my stops can't be hunted now, but they would be in a mechanical system. Yeah. I think I'm talking myself out of that idea.

Wow, two entries in one day? What's up with THAT?

Here's the latest refinement in the strategy. The basic idea is the same; using my research to determine when a price leg is most likely to be exhausted so that I can buy when downlegs are exhausted near support and sell when uplegs are exhausted near resistance.

But this time, I'm not going to anticipate the turn in the price leg. No more averaging into a loser. Instead, I wait for the price to pass the median projection into the "exhaustion zone" and then begin placing orders behind it, waiting for price to turn and start filling them when it's moving my way. Same scaling in deal with the orders; I use larger unit sizes as we move further into the exhaustion zone, so when price turns the largest order will be filled first at the best price. This way, as the orders are filled, the average entry price will be less and less affected at each new addition to the position.

The orders are 50 pips apart, and the last one is always outside the boundaries of the last day's price action (I use 1600 CST as the cutoff, not midnight CST which is where the candle boundaries are). This may cause a few of the orders to fill and even get stopped out due to volatility before the new leg actually begins, but that's ok.

In the setup below, I'm using 160 as a stop for the orders that are already open because that's just above the last major peak. Once I begin placing orders at 160 or above though, I'll have to use a higher point for those stops. 162.50 is the next major top so I'll be using that. So far none of the orders are filled of course, and it looks like I'll be placing a few more tomorrow.

Well the five sell orders on the EUR/JPY have all filled so I'm short 15 units (5+4+3+2+1) and in profit by about 130 pips. The upleg has ended so I've begun to place the buy orders behind the downleg now.

The refinements that I made are to begin the order placement starting right at the beginning of the leg (as opposed to waiting for an entrance into the "exhaustion zone") and continue until the last order is less than 100 pips behind the current price (as opposed to keeping the orders outside the range of the last day's price movemnt). This has the effect of increasing the number of orders, and consequently the size of the ones closest to price, giving me a greater probability of reversing direction with a profit each time.

So I've got 7 buy orders placed now for a total of 28 units so far (1+2+3+4+5+6+7). If price were to turn back up at this point and fill the last three (7+6+5) my short would be closed at a slight loss I think and I'd be net long 3 units. Because of this, I don't need the stop on the current short position anymore at 160.00 so I cancelled that. The stops on the buy orders are at 153.00.

I'm hoping price continues down enough for me to at least be able to place the next order for 8 units; that would lock in profits at the turn. Anyway, here's the current situation:

The overall account was down 2.4% this week due mainly to the average-down craziness in the last experiment and a decline in NZD/USD.

In addition to managing the EUR/JPY position, I'm now actively managing the NZD/USD carry position each day using a old contrarian technique from the 70's called Automatic Investment Management (AIM), developed by a guy named Robert Lichello. He wrote a book called "How to Make a Million Dollars in the Stock Market - Automatically!" Cheesy title huh?? : Anyway, Google it if you want to know how I'm managing the NZD/USD position.

In the last installment I talked about wanting to place at least one more buy order for 8 units. As it turned out, price declined enough for me to actually place two more, for 8 and 9 units. Since then, price has risen enough to fill both of those orders. Since I was originally short 15 units, and I just bought 17, I am now net long 2 units with a stop at 153.00.

My original short position of 15 units had an average price of 158.83. I bought 9 units back at 157 and 6 units at 157.50 for an average cover price of 157.2. This yielded a realized profit of 163 pips on the 15 units. The remaining 2 units were also bought at 157.50 so that's my average price now.

Price has penetrated below the median projected length of this downleg which was 60% of the prior LLHH upleg. So we entered the "exhaustion zone" for this leg and have since turned back up. However, we don't have an "official" upleg started until price reverses by at least 250 pips from the last extreme, and it hasn't done that yet. So since the downleg is still intact, I can't start placing sell orders yet.

If price continues down, I will continue placing buy orders behind it in preparation for the next upleg. That 153 stop is looking too close though, so I'll go down past a few more support levels and use 148 for the stop on any new buy orders. If it turns out that price has hit a bottom for now and is heading up (which is my guess), then my buy orders will start to fill, I'll begin placing sell orders behind the new upleg, and it's business as usual.

I'm liking this whole buy low and sell high thing. It's much more profitable than buying high and selling low. Huh. Go figure. Here's situation as of this morning Central U.S. time:

Ok, since price has moved up past the 250-pip reversal threshold, we now have a confirmed upleg. In the last update my position had already switched to a small 2 unit long. Now the 7, 6 and 5 unit buy orders have filled, bringing me to a total of 20 units long. There are still buy orders totaling another 10 units (4+3+2+1) above the current level, which means my potential long position could go as high as 30 units.

Since we are now in an upleg, I have begun placing sell orders behind it totaling 15 units. I'm putting the stops on these at 163.00. Since I have sell orders totaling 15 units below the price, I have cancelled 15 units worth of the stops that were down at 153.00. There is still a stop down there for 5 units and the unfilled buy orders still have stop instructions on them as well, so as always, my position is completely covered by emergency stop levels.

My median projected length for this new upleg is 110% of the last downleg which would bring it to about 162.00, so this is where the new "exhaustion zone" starts. The 110% figure is of course based on that big research project in my site archives. The last downleg was a HHHL which classifies it as a pullback in an uptrend. That's why it's colored cyan in the chart below. Even though it's a down LEG, I consider it part of a potential up TREND.

Here's the situation:

Here's a scalp I did today. Price was coming down from a LLHH upleg, so the median projected pullback for the downleg was 60%. These are nice setups because after a range expansion leg (LLHH or HHLL) there is a very low probability that another expansion leg will occur. So in this case, it isn't likely that price would penetrate the last low, and the most common outcome is that it moves up past the last high.

When I opened my system and did the zig-zag analysis, the price was already down in the exhaustion zone, so I immediately bought one unit. Then when it penetrated even further down I bought a second. If price had moved down past the prior low I would have gotten out. My expectation was that the price would move up past the prior high. Since this was a scalp, I didn't wait for that though. I just took profit when it was there. The whole trade was less than 10 minutes long.

This was the only trade I did in the session, and this brings me to 9 winning sessions and one b/e session. Since I made 5 pips on 2 base units, I count it as 10 pips. One more winning session to go before I've earned the right to increase the base position size.

No chart today since there isn't much of a change from the last one. The decline in EUR/JPY has triggered my 5-unit sell order, so now my net position is long 15 units.

Down 0.08% this week. Yawwwnnn.

My sell orders for 4 units and 3 units both filled, so my net position in EUR/JPY is currently long 8 units.

The price just managed to hit the 250 pip threshold for a reversal, so the upleg is officially ended and a new downleg is in progress. Since the upleg ended, I cancelled all unfilled buy orders. Normally, I would have begun placing a new sequence of buy orders starting at the last peak, but I'm modifying the system.

The new rule is that I will not place the sequence of orders behind the current leg until the leg has surpassed its median projected length.

These median projections are the edge with which I'm supposed to be experimenting, but the way I was doing it before ignored these projections. Given this new rule, I will not commit to placing buy orders in preparation for the next upleg UNTIL this current downleg shows that it's extended into the exhaustion zone. Since I always start the sequence at the last peak/trough, this rule has the effect of increasing the size of the closest order as well as moving large sell orders higher and large buy orders lower which is where they should be. It is essentially a whipsaw reduction rule.

Here's the current situation:

The continuing decline in EUR/JPY has now triggered my 2-unit sell order, lightening my long position to 6 units. There is still a 1-unit sell order at 156.00 and a 5-unit stop loss at 153.00.

I have begun weekly tracking of the equity curve for just this portion of the account in addition to my normal week to week tracking of the entire account equity. The three strategies I run in the account continue to be:

1. This EUR/JPY position, managed using statistical projections for leg length
2. Scalping sessions on EUR/JPY
3. Long term low/no leverage NZD/USD position managed with Lichello's AIM concept

EUR/JPY went down far enough to trigger my 1-unit order so now I'm long only 5 units. The downleg has ended and we are in a confirmed upleg now. The last downleg was a LHLL which makes it a trend leg, so the median pullback is 80%. Price has moved above that, so we're in the exhaustion zone.

This allows me to place sell orders below the leg, and I've started that process with the largest order being for 6 units so far. Since I have sell orders below the price now, I have cancelled my stop loss on the long position at 153.00 My stop for the sell orders is up at 165.00.

Here's the current situation:

I had another successful scalping session this morning. That brings me to 10 winning sessions and one break-even session at this position level, so I've finally earned the right to trade larger!

I've been a little concerned with my scalping method though because it involves an element of averaging down. So after falling behind a bit in today's session, I decided to try a variation of the technique I use in my longer term position management with EUR/JPY.

So as a leg moves beyond it's projected exhaustion level on the 5-min chart, I will place an order behind the price (if price is dropping I place a buy order above and vice-versa). I don't place a series of orders like I do in the long term position because in a scalping strategy I want to close out the position. What I do instead is just place the last order in the series, using the same unit size that I would have used if I had placed a series.

So for example in this session, price had dropped past an exhaustion level. So starting at the last peak, the normal series would have been buy orders for 1, 2, 3 and 4 units separated by 10 pips. I actually did place all four of these orders, but cancelled the first three when I realized what I actually should be doing. So my only order was for 4 units at the lowest spot.

If price had continued to drop, I would have cancelled that order and placed a 5-unit order another 10-pips further down into the exhaustion zone. If the 4-unit order filled, and then price continued down, I would continue to place larger orders behind the price, cancelling any unfilled orders as I placed each new one.

I'll use an emergency stop beyond several S/R levels, and take profit when the leg reverses in my direction and then begins to lose steam or enters a new exhaustion zone.

Here's this morning's session: