Caution: Math Geek at Work | Page 13 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Murrey Math indicator is very useful in Forex 69 replies

Murray Math Levels 30 replies

Murray Math 13 replies

Turbo's Murry Math 23 replies

Attachments: Caution: Math Geek at Work
Exit Attachments

Caution: Math Geek at Work

  • #241
  • Jun 15, 2008 8:09am Jun 15, 2008 8:09am
Quoting capitalist88
No, I think mixing the two time frames just confuses the issue, so let's keep it as simple as possible. The question is essentially whether the short term 50-hour correlations are meaningful or just the result of randomness. In order to do this, you need to compare several independent periods to each other.

The tool I would pick is the coefficient of variance (COV). This is just the standard deviation of a set of data divided by its mean, so it gives you a dimensionless measure of how "consistent" the data points are. You could take correlations from 10 periods.

The ideal case is that all the correlations are high, and tend to hover right around the same values. So suppose you get results where the mean is very high, say around 90% and the STDEV is low, like around 10%. This would give you a COV of 0.11. I would say any COV of 0.10 or less would indicate that there is SOMETHING (no idea what) non-random which is causing a CONSISTENTLY high correlation.

Now I don't know which would work better for you; to take several periods of 50 hours each, or to break up one 50 hour period into several smaller periods.

Either way, you can use COV to test to see if the correlations you're seeing are just the result of randomness or not. This is a very basic "back of the envelope" kind of test, and a real statistician would probably laugh at me, but it's quick. One thing to remember is that in order for COV to give meaningful results, ALL OF THE DATA POINTS MUST HAVE THE SAME SIGN.

Hope this helps a bit.
Thanks! The results are listed below. Base condition was 25 Pip TP/SL with 4 hour timed exit using over 1.5 years worth of 15 minute candlesticks. In subsequent tests I ran with a COV using 5, 10, 20 and 50 correlation points. I then add a condition to check that for positive correlations the COV was > .1 and negative correlations the COV < -.1

Base condition:
Pips: 7,488
Win/Loss: 0.665946

.1 COV, 5 bar conf:
Pips: 7,536
Win/Loss: 0.667389

.1 COV, 10 bar conf:
Pips: 7,415
Win/Loss: 0.670034

.1 COV, 20 bar conf:
Pips: 6,270
Win/Loss: 0.668421

.1 COV, 50 bar conf:
Pips: 3,780
Win/Loss: 0.651072

The drop from 20 bar to 50 bar is interesting. Also, just for grins, using my money management approach options 2 and 3 will turn a $10,000 investment into > $400,000 during this testing period. Of course that is just theory, making it work in reality is another thing:-)

Dan
.
  • #242
  • Jun 15, 2008 9:43am Jun 15, 2008 9:43am
Quoting redpine
Thanks! The results are listed below. Base condition was 25 Pip TP/SL with 4 hour timed exit using over 1.5 years worth of 15 minute candlesticks. In subsequent tests I ran with a COV using 5, 10, 20 and 50 correlation points. I then add a condition to check that for positive correlations the COV was > .1 and negative correlations the COV < -.1
Hmm? Those conditions look reversed. You want the COV of the correlations to have an absolute value that is as low as possible.

Anyway, at a 2/3 win rate and a 1:1 risk/gain ratio, your expectancy is already 2/3 - 1/3 = 1/3 or about 33% per trade. This is extremely high. If you can really get anything even close to this in a live environment, you'd be better off spending your time sipping Mai-tais on a beach than trying to improve the performance, LOL.
  • #243
  • Jun 15, 2008 10:56am Jun 15, 2008 10:56am
Quoting capitalist88
Hmm? Those conditions look reversed. You want the COV of the correlations to have an absolute value that is as low as possible.

Anyway, at a 2/3 win rate and a 1:1 risk/gain ratio, your expectancy is already 2/3 - 1/3 = 1/3 or about 33% per trade. This is extremely high. If you can really get anything even close to this in a live environment, you'd be better off spending your time sipping Mai-tais on a beach than trying to improve the performance, LOL.
Opps. That's why your the math geek. Software people like me tend to fill in the blanks and occassionally get it wrong. I'll try it with an absolute correlation measure. Thanks.

Someday I hope to be too busy deciding how to spend my money

Dan
.
  • #244
  • Jun 16, 2008 8:32pm Jun 16, 2008 8:32pm
The EUR/JPY continues to rise against my short position. It went up enough today to make room for a 10-unit sell-stop order which I've placed. That's all...
  • #245
  • Jun 18, 2008 10:14am Jun 18, 2008 10:14am
EUR/JPY continued up a little more today to over 168.00. Using my 100 pip buffer for placing orders, this gave me room to place the eleventh sell order in this sequence at 167.00.

Also, since the 9-unit order filled (did I post that?) a few days ago, there was another gap that I had to fill. Last time I filled a gap, I replaced the 1,2, and 3 unit orders with a 7-unit order at the higher price. This time, I used the 4 and 5-unit orders to fill the 9-unit order gap.

So before the gap filling, my sequence looked like this (sell orders placed at 50 pip intervals): 4,5,6,7,8,GAP,10,11.

Now it looks like this: 6,7,8,9,10,11.

Same amount of units present in the sell orders, but the average price at which I'll be selling has gone up. Again, that's the point of the method. Always increase the selling price and always decrease the buying price when the opportunity arises.
  • #246
  • Jun 19, 2008 7:19pm Jun 19, 2008 7:19pm
EUR/JPY came down a bit and hit my 11-unit sell order at 167.00. My short position now is 95 units. My average selling price for the entire life of the strategy is about 160.00 and the average buying price is 158.00, for a total spread of 200 pips.
  • #247
  • Jun 22, 2008 9:32am Jun 22, 2008 9:32am
Wow, up 1.1% this week. Better than down 4%. The recovery in NZD/USD is what drove the profit. The EUR/JPY position is still in drawdown mode.
  • #248
  • Jun 25, 2008 10:53am Jun 25, 2008 10:53am
The EUR/JPY continues its upleg, passing through 168.50 a couple of minutes ago. This gave me the 100-pip buffer space to place a 12-unit sell-stop order at 167.50. Also, since my 11-unit order had already filled, there was another gap in my order sequence, which looked like this starting at 164.50:

6,7,8,9,10,GAP,12

So as usual, I put the 11-unit order back in and compensated by cancelling enough orders further back to cover the volume. The sell order sequence now looks like this, starting at 165.50:

8,9,10,11,12

Notice that the order density is rising as larger orders are getting squeezed into a smaller space.
  • #249
  • Jun 26, 2008 9:38pm Jun 26, 2008 9:38pm
Greetings from DrawDownLand!

EUR/JPY still rising, and broke some long term resistance levels too. Placed another sell-stop for 13 units at 168. The sequence is now:

8,9,10,11,12,13
  • #250
  • Jun 29, 2008 10:43am Jun 29, 2008 10:43am
Ok, we've got a confirmed reversal from an upleg to a downleg in EUR/JPY so it's time to post another chart.

The two past down days have filled my 13,12 and 11-unit sell-stop orders, bringing my total short position to 131 units at an average price of 164.59. The 10,9 and 8-unit orders are still open.

The last upleg was a trending leg which went from a higher low to a higher high (HLHH). Trending legs have a median pullback of 80% based on the research posted at my site, so this puts the exhaustion level for the new downleg down at 163.27. I can't begin placing buy orders behind this leg (meaning buy-stops above the price as it declines) until price goes below this level.

Also, if that happens, I will not be placing a sequence of orders. Instead I will be be "collapsing" the normal initial sequence into one combined order so that the average buy price is lowered as much as possible. This "collapsing" technique is what I use whenever I'm placing orders that reduce my position and thus my risk. Since I'm currently short, any buy orders will reduce my risk, so I'm happy to consolidate them down to the lowest point possible.

But that's all just speculative for now, since I don't know if the price will get low enough for me to begin placing buy orders yet. Anywyay, here's the chart of the current situation:

Attached Image (click to enlarge)
Click to Enlarge Name: position.JPG Size: 86 KB
  • #251
  • Jun 29, 2008 10:48am Jun 29, 2008 10:48am
The account was pretty flat this week, with a loss of 0.2%. The calm before the storm?
  • #252
  • Jul 2, 2008 10:10am Jul 2, 2008 10:10am
A couple of things have happened in the EUR/JPY position.

My 10-unit sell order filled, bringing me to a total short position of 141 units. Also, the direction has flipped again. We're now in another upleg. The last downleg was a trend retracement leg (HHHL) which has a median reaction of 110%. This puts the exhaustion level for the new upleg at 169.80. So I won't be placing any more sell orders unless we get up above that level.

Since we're finished with the sequence of orders that I built behind the last upleg, I have cancelled the two remaining open sell orders for 9 and 8 units. If this new upleg gets above the exhaustion level, I'll start building the sequence from 1 unit again.
  • #253
  • Jul 6, 2008 9:37am Jul 6, 2008 9:37am
Down 1.0% this week. But I sense forces gathering for a rebound...
  • #254
  • Jul 14, 2008 4:58pm Jul 14, 2008 4:58pm
Will it never end? Down 1.0% again last week.

The EUR/JPY, while it has gone up against my short position some more, still hasn't gone up past the current exhaustion level. So I haven't placed any more sell-stop orders below this upleg yet. Not that I really want to... my margin is getting a bit thin at this point. Bleh.
  • #255
  • Jul 15, 2008 12:56pm Jul 15, 2008 12:56pm
Yay! A new downleg has started on EUR/JPY based on the 250-pip reversal threshold that I use.

The upleg that just ended was a trend extension leg (HLHH) which has a median reaction of 80% of its length. This put the exhaustion level for the new downleg down at 166.80, and the price has already gone below that level. This allows me to begin placing buy-stop orders above the leg.

The leg currently reaches from 169.68 to 166.48. I start the sequence at 169.50 and stop at 167.50 to maintain the 100-pip buffer for order placement. This would normally mean that I place 5 orders starting with 1 unit at 169.50 and continuing with 2,3,4 and 5 units each 50 pips. However, since buy orders will reduce my risk, I "collapse" all of these orders into one large one at the lowest possible price. So I placed an order to buy 15 units (1+2+3+4+5) at 167.50. If price continues south, I will keep moving this buy order down and increasing its size according to the sequence.

Note that this buy order is above the average sale price for my short position of 141 units. However, it is 50 pips below the last individual sell order that filled for 13 units at 168.00. So those 13 units have a locked-in profit of 50 pips. This represents a net improvement over my previous position.

Here's the chart.

Attached Image (click to enlarge)
Click to Enlarge Name: position.JPG Size: 84 KB
  • #256
  • Jul 16, 2008 10:37am Jul 16, 2008 10:37am
Ok, a couple of things have happened.

The current downleg on the EUR/JPY has reached below the previous low, so this leg is a range expansion leg (HHLL) which ends the uptrend composed of the last three legs (marked with blue lines below).

Also, the downleg has moved by another 100+ pips, giving me room for two more orders in the sequence; the 6-unit and 7-unit orders. Of course, since I'm using a single "collapsed" buy stop order, I just add these to the existing 15 unit order and move it down. So the new order is a buy-stop of 28 units (15+6+7) at 166.50.

This locks in even more profits on the last few sell orders that executed during the uptrend retracement leg. Here's the chart:

Attached Image (click to enlarge)
Click to Enlarge Name: position.JPG Size: 79 KB
  • #257
  • Jul 20, 2008 3:28pm Jul 20, 2008 3:28pm
Volatility abounds!

EUR/JPY has reversed it's downleg and started a new upleg. The 28-unit buy order I had at 166.50 filled of course, bringing my net position down to 113 units short.

This upleg has passed its exhaustion level already, so I've started to place sell orders below the leg. The leg has moved enough to give me room for 7 orders so far.

Here's the current situation:

Attached Image (click to enlarge)
Click to Enlarge Name: position.JPG Size: 82 KB
  • #258
  • Jul 20, 2008 3:31pm Jul 20, 2008 3:31pm
Down 0.5% last week. Ho hum...
  • #259
  • Jul 23, 2008 7:39pm Jul 23, 2008 7:39pm
hi capitalist,

you got a nice thread going here and I appreciate you sharing your strategy with us here.

I suppose you like the idea of setting your orders for the days ahead and then monitor the situation once a day.

I just wonder if you should place your first sell orders higher on the chart and you buy orders lower so your average price becomes higher and lower.
Your profit will increase as well.
Sometimes you will miss out on some trades towards the middle of the swing, but those are more loss prone anyway.

I also would trade this strategy without a stop loss on more pairs, certainly after trialling proves the strategy is viable.

In my opinion the best thing would be to buy more at the bottom of the swing and sell those at the top and vice versa.
But here you would have to monitor your trades on some days when price comes in your sell/buy zone say the top and bottom 10% of the swing.

If its a uptrending pair, I would want to buy in the bottom 20% and sell in the top 10%.
The difference is the fact that price might go through the earlier topmost price and your losses will accumulate.

Unfortunately I haven`t had a chance yet to verify my statements, but before doing so I thought it might be a good idea for you to give your opinion.

thanks and take care.
  • #260
  • Jul 24, 2008 11:29pm Jul 24, 2008 11:29pm
Hi yonnie!

I think I'm already doing what you suggested. My largest sell orders are always higher than the small ones, and vice-versa for buys. Take a look at the sell orders in post #257. The 7-unit order is up at 168.50 and the 1-unit order is down at 165.50.

So the median price of this series of orders is 167.00. But if all 28 units were to fill, my average sale price would be higher than that. It would be:

(7x168.50 + 6x168 + 5x167.50 + ... + 1x165.50) / 28 = 167.50.

Once I get more capital to work with, I might expand it to other pairs as well. I would just have to do the same sort of analysis as I did for EUR/JPY (see the research notes archives on my site) to get the median reaction levels.

Also, I agree on the stops. I don't set stops anymore for this method. I forget when it was, but there's a post where I talked about eliminating the stops.


Quoting yonnie
hi capitalist,

you got a nice thread going here and I appreciate you sharing your strategy with us here.

I suppose you like the idea of setting your orders for the days ahead and then monitor the situation once a day.

I just wonder if you should place your first sell orders higher on the chart and you buy orders lower so your average price becomes higher and lower.
Your profit will increase as well.
Sometimes you will miss out on some trades towards the middle of the swing, but those are more loss prone anyway.

I also would trade this strategy without a stop loss on more pairs, certainly after trialling proves the strategy is viable.

In my opinion the best thing would be to buy more at the bottom of the swing and sell those at the top and vice versa.
But here you would have to monitor your trades on some days when price comes in your sell/buy zone say the top and bottom 10% of the swing.

If its a uptrending pair, I would want to buy in the bottom 20% and sell in the top 10%.
The difference is the fact that price might go through the earlier topmost price and your losses will accumulate.

Unfortunately I haven`t had a chance yet to verify my statements, but before doing so I thought it might be a good idea for you to give your opinion.

thanks and take care.
1 trader viewing now