Caution: Math Geek at Work

Scott, thank you very much again. You gave me a new perspective on how to measure performance. And the link about posting images was cool! So let's continue the fight. Good trading. I will be following your thread with great interest. Keep up the good work.

The price spiked down far enough to fill my 12 unit order, so now I'm short 86 units of EUR/JPY. Price then moved up again by more then 100 pips, so there was a "gap" where that 12 unit order used to be. So I put in another 12-unit order there, and compensated for that by eliminating the lower priced orders for 1,2,3,4, and 5 units. Just another instance of moving the orders to more advantageous levels when there's a gap in the sequence.

Today's downturn in EUR/JPY caused two more sell orders to fill. The 12 unit and 11 unit orders filled bringing my short position up to a total of 109 units (the largest position ever so far) with an average entry of 160.72.

The downturn hasn't quite reached the threshold where it becomes a direction reversal yet though, so we're still in the upleg.

OK, I'll now become a trading heretic.

Since my stops are always like 1000 pips away, and getting stopped out would seriously mess up my average buy and sell prices, and since my leverage is generally very low, I am no longer using stop orders.

Let loose the hounds!!!!!

Hi Scott,

With an xK standard account, MHE (my humble experience ) points out that a position size of 0.0x lots needs no stop at all. I consider the risk of this size NEAR zero. Sure, some people will never agree with that. But I think we can establish a profitable modus operandi both using stops or not. Good trading.

Hmm, that's a pretty useful rule of thumb. My EUR/JPY position right now is just slightly over 0.01x lots by that measure.

My NZD/USD position (which I don't discuss much here) is larger, but because I always hold that long for the interest income, and the USD is my home currency, I don't need a stop on that. The risk is completely limited by the position size.

The EUR/JPY position was profitable this week, but my NZD/USD position lost value. The overall effect was a loss of 0.7% of total account value.

Here's the EUR/JPY position performance.

Employed capitalist is employed.

Well this is a bit more like it. The EUR/JPY is now in a downleg and three more of my sell orders filled for 10,9 and 8 units. My short position is now up to 136 units. There are two more unfilled orders for 7 and 6 units.

The upleg that just completed was a HLHH trend leg. This type of leg has a median pullback reaction of 80% which determines the exhaustion level for the current downleg. I won't be placing buy orders above the price in this downleg unless price moves below that level.

If and when I do start placing buy orders, they will have the effect of reducing my position, so they are "profit taking" type orders. That's the situation where I use the "order collapsing" technique. I'll start counting for the order series up at the last peak as usual, starting with 1 unit at 165.00, 2 at 164.50, etc. But I won't actually place the orders up there. When the price is below the exhaustion level, I'll add up all the orders up to a point below the breakeven point and combine them into one order, then I'll start placing orders as usual below that.

This keeps my average buy price as low as possible. There's an example of this a few pages back in this journal. Anyway, here's the current situation:

Price moved far enough down today to trigger the 7-unit sell order, so now my short position in EUR/JPY is 143 units at an average entry price of 160.99. No reason for a chart today.

Still treading water as the account gained 0.3% this week. The EUR/JPY position open equity is below.

Howdy,

Nothing much to report. EUR/JPY went up to a point just 2 PIPS below the 250 pip reversal threshold that I use to define a change in direction from one leg to another. Then it came back down from there. So there is no change in leg direction, we're still in the downleg and price is just kind of drifting sideways.

w00t!!!

It's 3:00 AM. I'll post more on this w00t tomorrow...


Merlin, if you read this... Love the new smileys!!

The downleg on EUR/JPY has crossed below its exhaustion level so I can start placing buy orders. Since buying will reduce my risk, I use the "collapsing" technique on the order sequence. There is enough room for 10 orders, so I collapsed that sequence down into one order for 55 units (sum of the sizes from 1 to 10). This order is in profitable territory, so I've locked in profit on about 1/3 of my short position. w00t!

Even in spite of the drop in my NZD/USD carry position, the entire account was up a whopping 4.0% this week. Of course the driving force was the nice drop in EUR/JPY covered in the last couple of posts. Here's the open equity of just that position management portion of the account:

A few things have happened since the last post.

1. In addition to the 55 unit buy order I discussed in the last update, EUR/JPY went down far enough to make room (100 pip buffer) for the next order for 11 units (remember the 55 unit order was a consolidation of the sequence from orders 1 to 10).

2. I realized that the portion of the account dedicated to this strategy has hit the point where I was supposed to increase the order size by 10%. So I had to add two additional buy orders to bring the existing two up to the new order size.

3. Price turned back up and crossed the 250 pip reversal threshold, so now we're in a new upleg. All four of the buy orders have filled.

4. I had to make a tiny market buy order (which I messed up twice) in order to make my current position an even multiple of my new basic order size. So I started out short 149 units of the old size. I bought 66 units of the new size, but my new position isn't short 149 - 66 = 83 units. Because of the new order size I'm dealing with, I'm now short 68 units of the new size.

Get it? This order size adjustment stuff will happen whenever the profit gets high enough to justify a change. Anyway, I have no open orders now, and the exhaustion level for this upleg is up above 165, so here's the current situation:

Down 2.9% this week. Not much else to report. Bleh.

Hi there

Sorry to bump into your journal uninvited...given your title of math geek I wanted to ask you if you ever experimented with non-linear,non-parametric regression and copulas in regards to fx trading.

I'm trying to sink my teeth into this stuff but given the fact that I don't know that much math, it's going a bit slow.

Since it's the weekend, I thought I'd put a little blurb in here about the underlying idea behind this "position management" method of trading the EUR/JPY.

The basic concept is the same one a retail store owner uses to make a living. Buy stuff at bargain prices in bulk, and resell it when the demand is higher. Over time, the volume of these transactions combined with the spread between the average buy and sell price creates a profit.

My position management system is just a mechanical set of rules (based on some statistical research) that tend to push my average buying price down and my selling price up. As long as the spread between the two stays positive (buys lower than sells), the more volume I do, the more profit I get.

So far, I have done 69 individual trades since 2/28/08 with this system. I downloaded the transactions from Oanda and used Excel to calculate my average buy and sell prices over time. So far so good! Here are the results:

Never used copulas, but maybe non-parametric methods woud have been useful with some research I did using ordinal level variables. For example I was once looking at the relationship of prices to economic releases and how the releases compared to the estimates and the previous month's actual value, so I had to use an ordinal level variable to describe this. I didn't use any formal non-parametric methods to analyze the data though.

I've gotten away from predictive models though and am now just adjusting my positions based on where price is now, as opposed to where I think it's going to go.


For more info on my actual research, check the archives of my site, www.market-geeks.com