Daily dow system (cantor should put me on commision!)

I have a system designed to trade the daily dow contract with Cantor Index.

The advantage of my system is that it the entry points are determined for the following day after the US close (around 21:30 GMT) and can then be entered for the following day's contract.

This means entry orders can be made in a few minutes max.

The system takes long or short positions and if neither entry is hit will not trade in that day.

The position taken is settled at the end of each day against the official closing price of the Dow.

I have tested this system with a varying betsize depending on volatility but as it does not have any stops it can be viewed as risky.

Performance since 01/01/96 goes like this:

2775 trading days 1760 winners, 603 losers and 412 days it didn't trade

Average winner = 82.5 points average loser = 49.6 points

Max winner = 460 and max loser = 453.85

In just over 11 years it has bagged 115,000 points off the DJIA around 10,450 per year

Even at a conservative level of risk, say 200 point loss was 2% of account £100,000 would bet £10pp and have made around £100,000 each year without reinvesting and upping the betsize.

This of course all sounds too good to be true and as I am no maths genius I am trading it live waiting for it to go wrong!

Signals this year have been

Sell Buy
03-Jan-07 12536 12390 61.48
04-Jan-07 12549 12400 0.00
05-Jan-07 12553 12408 -9.99
08-Jan-07 12472 12324 0.00
09-Jan-07 12474 12334 0.00

So it is +51.49 this year.

Today's signals are S@12,474 and B@12,334.

As always even if this system has potential it will be risk control and betsizing that will determine if it become viable or not.

I have tried this system on other markets to good effect and currently trade a slightly optimized version of what I'm writing about but even in it's basic form with decent money control I think it can be a goer.

Not sure what the purpose of this post was but if anyone is interested in the daily signals let me know and I'll post them or email them.

Might be fun to track the progress of an auto system.

Stephen McCreedy
[email protected]

Hello Stephen, sounds a bit like one of the systems I use to trade the daily Dow. I would be interested to see your daily signals......

Nick

Dow orders 10/01/07

Brought forwars since 01.01.07 51.49 points.

Yesterday the orders were B@12,334 and S@12,474 neither were hit so we are still on 51.49 points profit so far this year.

Today's signals are B@12,348 and S@12,485

Don't forget this is a system and you just have to let it ride, one good or bad trade won't prove it either way only months of consistent following will show how worthwhile it is.

Money management is upto you but as an example I trade at a size that the a loss equivalent to the distance between the two orders would equate to losing 5% of my account.

Don't forget this system has no stops so much more than this could be lost but in today's instance the market would need to fall or rise 137 beyond the entry ie 12,211 or 12,622 around 200 points from last night's close, for me to lose 5% of my account, this is risk that suits me but isn't going to be right for everyone.

Stephen McCreedy

win lose ratio max win max lose
1996 6798.3 173 42 80.5% 131.00 -112.55
1997 6854.03 158 60 72.5% 235.00 -453.85
1998 9274.18 164 52 75.9% 316.00 -376.93
1999 17570.41 173 45 79.4% 347.00 -134.08
2000 21340.46 187 33 85.0% 460.00 -382.23
2001 12914.34 150 50 75.0% 439.00 -269.75
2002 8376.71 136 81 62.7% 369.00 -244.88
2003 5814.45 140 73 65.7% 200.00 -197.32
2004 7217.58 145 63 69.7% 187.00 -85.76
2005 9070.79 165 53 75.7% 195.00 -109.60
2006 10008.45 168 50 77.1% 198.00 -136.61
2007 51.49 1 1 50.0% 61.48 -9.99

Total 115291.19

Here are stats for 11 years and a few days of the system.

Make for nice reading, only time will tell how tradable the signals are

Stephen McCreedy

That didn't format so well so here is a jpeg of the results.

Left column immediately after the year is the net points profit made on the DJIA from entering orders and hanging on till close.

As you are buying low and selling high the maximum gain is set at the distance between the two orders and the maximum risk to longs is theoretically 100% of the market and for shorts infinity or more realistically 100% of your account!

As said earlier position sizing is everything

Stephen McCreedy

smccreedy said:

That didn't format so well so here is a jpeg of the results.

Left column immediately after the year is the net points profit made on the DJIA from entering orders and hanging on till close.

As you are buying low and selling high the maximum gain is set at the distance between the two orders and the maximum risk to longs is theoretically 100% of the market and for shorts infinity or more realistically 100% of your account!

As said earlier position sizing is everything

Stephen McCreedy

Click to expand...

interesting.. may i ask what data source you are using?

i ask because i am yet to find a daily system at yours that sells high and buys low and still turns a profit when exiting at the end of the day of entry.

ie, you can buy at the open +0.1% 0.2% 0.3% etc, and every single level is long term profitable (depending on spread requirements) which is why im surprised you can sell at such levels and still turn a profit.

fc

I use yahoo which can be a pain for FTSE as opening equals the previous close but with the DJIA I've found the data pretty accurate.

The reason I have kept a healthy cynicism myself is that of course trades that according to data have been hit might not actually get filled.

This is why the only way to see is to trade it.

I can report in the few days this year I have not had a problem getting my orders on the system as they are far enough away from the market and my real trading has performed perfectly in line with the system.

I have a more profitable system through changing some parameters but the problem is that getting the orders on his tough as they are too close to the market. This means they are entered real time not in advance so might be suitable for some people but not others. With this system I enter orders around 8am for the Dow session that day based on closing date the previous night.

A good example is last night where my personal system (not the one I'm talking of here) netted 29.4 points and this one didn't trade and on Monday when the tighter system made 49.94 and this system didn't trade.

The flip side is sooner or later you'll record a loss with the optimised system that this system I have been quoting doesn't trade as the order didn't get hit.

The key point is really that with the instrument I trade I can get the orders in before the market.

Yesterday is a good example.

My personal system went short at 12446 and this was triggered around midday London time.

Only time will tell how easy it is to replicate in the real world the nice looking system I have on paper. I feel it won't be a problem with the system I'm publishing here, it is being a pain on the FTSE if I don't come in to work at 9pm to get orders on for next day and is being hard with the optimised version to get orders on.

Still I'd rather work with something I know has worked in the past I'd I can replicate reliably than try to second guess something more profitable that can't be copied in the real world.

Stephen McCreedy

Sorry to drone on about this but in ref to your post chinos.

With this system as you are buying at low levels and selling at high levels the case of orders not being hit couldn't harm the account.

This is because if I want to buy low say 12,500 the data says I could but in reality the market only trades there on the open before falling to 12,400 that is a trade that would lose money. Same applies to selling before I rally occurs. This means any trades that system shown as being hit that the market couldn't hit are automatically losers. In effect this means the system in reality is maybe a little safer than the paper system.

Hope that makes sense. The convers is true of my FTSE system which buys breakouts and sells downward breakouts, this means the system might get a fill when the real world doesn't and this would mean some profits in theory won't be made. With the Dow it means some losses in theory won't be made.

Hi SM
Yahoo data is inaccurate as the way they calculate the intraday high and low means there can be a discrepancy with the genuine market also their opening price may not be a tradable price at all

smccreedy said:

I use yahoo which can be a pain for FTSE as opening equals the previous close but with the DJIA I've found the data pretty accurate.

The reason I have kept a healthy cynicism myself is that of course trades that according to data have been hit might not actually get filled.

This is why the only way to see is to trade it.

I can report in the few days this year I have not had a problem getting my orders on the system as they are far enough away from the market and my real trading has performed perfectly in line with the system.

I have a more profitable system through changing some parameters but the problem is that getting the orders on his tough as they are too close to the market. This means they are entered real time not in advance so might be suitable for some people but not others. With this system I enter orders around 8am for the Dow session that day based on closing date the previous night.

A good example is last night where my personal system (not the one I'm talking of here) netted 29.4 points and this one didn't trade and on Monday when the tighter system made 49.94 and this system didn't trade.

The flip side is sooner or later you'll record a loss with the optimised system that this system I have been quoting doesn't trade as the order didn't get hit.

The key point is really that with the instrument I trade I can get the orders in before the market.

Yesterday is a good example.

My personal system went short at 12446 and this was triggered around midday London time.

Only time will tell how easy it is to replicate in the real world the nice looking system I have on paper. I feel it won't be a problem with the system I'm publishing here, it is being a pain on the FTSE if I don't come in to work at 9pm to get orders on for next day and is being hard with the optimised version to get orders on.

Still I'd rather work with something I know has worked in the past I'd I can replicate reliably than try to second guess something more profitable that can't be copied in the real world.

Stephen McCreedy

Click to expand...

i suspected you were using yahoo data.. and i dont mean to **** on your bonfire, but the highs and lows in there historical data are hugely flawed for every single day

eg look at yesterday's data

Date Open High Low Close
9-Jan-07 12,424.77 12,516.66 12,337.85 12,416.60


the market didnt get anywhere near 12516 at any time..

try testing your system on SP500 or FTSE100 and see if you have any luck. chances are that you wont and it will be hugely negative.

just thought i would save you some testing time, as ive been down this route before.

fc

Data integrity is the largest reason I have to be sceptical with this interestingly though my OHLC data for yesterday is

12424.77 12466.43 12369.17 12416.60 which I see as pretty in line with what I witnessed live.

The close is 100% accurate and important as what my system is settled against so no problem there.

The open is irrelevant to my system in all calculations and has no bearing so no problem there.

The high and low in my experience of monitoring it and now trading it are that there is little variance. Some times it works for me and sometimes against.

I have discarded trade where highs and lows = the open in the past to take in to account un tradable trades and there was little difference.

Also more often than not when I trade my trade at times is wildly out of pocket and scaring me before settling back to close in the range. This means that even if highs and lows are not accurate as because my entries are often well below/above the high/low they don't really have much bearing.

I concede some trades I bank on entering won't get entered and that is why I am trading this with 100% accuracy to see how it goes.

Incidentally the FTSE system works well on Dax and S&P too.

If anyone was to ask for 'proof' of orders getting filled that would be fine but it's so early in the year it proves little. all I can say is so far system and actual match 100% on DJIA.

Thanks for your input everyone.

Stephen McCreedy

I think the bad Yahoo data must cause a lot of people to lose money.

Best to test on accurate futures data to see if something really works.

In this system it won't lose me money on a trade it might mean system winners don't materialise which might cause the system on the whole to lose.

There aren't too many losers in this system it's got quite a high win ratio so although the loss of some winners will ruin the system I'm still unsure if it would result in a loss overall.

Easiest way to see is to trade it.

As I have said all signals have been followed so far this year today's orders are in and we'll wait and see the result.

I am working on a website to publish orders and can record fills so hopefully we'll all learn something.

I am interested to why my yahoo data is different to Chinos' though?

Stephen McCreedy

I am interested to why my yahoo data is different to Chinos' though?

Click to expand...

you are looking at the chart page with yesterdays data chino is looking at the page with historical data which you would take your data from

Are we allowed to ask how you calculate the levels for the entry orders? (You may of course wish to keep this 'secret'). Whilst I take your point, with regard to the closing level being 100% correct on Yahoo, if the entry levels are calculated (on the historical backtest) using flawed data then does it not throw a spanner in the works so to speak?

Very interesting thread by the way.

I would agree with what you have written with regards to the testing. You cant beat testing it in real market conditions. Given the number of points that it seems to make in the backtest it would become pretty obvious if the actually performance was drastically reduced.

Watching with interest here.

Steve.

right, here we go.. a basic system to show the problems with yahoo data

enter each day at the close, and take a profit at 1% of the index value..

eg last night, we would have entered a long at 12,416 with a target at 12,540

on the yahoo data this makes 57,000 points.. with futures data it loses 7,000 points

the reason why this is the case is that yahoo stretches the highs and lows to places that the market never traded to. because the close is where the market traded, it is obvious that by shorting from an artificially high place and closing in a real place that it is going to be profitable.

try testing your strat on the futures data i have supplied in my file. i would be surprised if reversing your entry rules wasnt profitable.

the 2 images i attach as well are of the yahoo data first, and then the futures data.

Big difference, no?

fc

Sure does, I take your point as you're getting filled at imaginary levels.

As you state if you could sell at levels that didn't exist and cover at the close it's a no brainer.

What I am saying is in my system the orders are not that far from the previous market close and in my limited time following this system it hasn't been a problem yet.

If I was selling the top tick on yahoo and covering at the close and buying the false low and exiting on close the system would of course be nonsense.

However as I've maintained all along my interest is sufficient to trade real time along my system and see what variance exists. I am sure there will be some and to the detriment of it but how much remains to be seen and I am eager to find out.

Thanks for your input in to the debate though, very useful and definitive.

What interests me is that the FTSE system and S&P system and Dax system all work alright but by no means as well as the dow.

Is the data more reliable on these markets?

I've always hated the DJIA, the way it sets the open and close and the way it is weighted by share price not market cap makes for a crazy system but like any data it can be traded and the volatility is certainly there to encourage people.

Stephen McCreedy

Hi just saw the data you sent.

Any idea why the DJIA is so false on yahoo and is the FTSE more reliable?

Thanks

Stephen McCreedy

because in the olden days (pre 1960 i believe maybe even earlier), the highs and lows for the day were calculated assuming that all the stocks in the Dow hit their highs and lows at exactly the same time.

for some weird reason yahoo decided to adjust all the recent data to be calculated in the same way, rather than restate the old data.

it is an almighty pain in the ****.

for all the other indices that yahoo has data for, it appears to be correct give or take a couple of points here and there..

fc

Like I stated earlier, isn't the Dow an almighty shame of a benchmark!

Have quickly tested the data and it shows a profit from my buy high system and loss for the sell high system (The one I've published acutal orders for on here) That isn't conclusive as haven't taken the time to really dig in to it.

As I said before I was waiting for mistakes to crop up and feel better they have.

I want a decent real system not a crazy pie in the sky system.

What is nice is that it reinforces my other system works and I have plenty of faith in the workings of it but was worried about data, now it seems data is not as much of an issue on the FTSE system I am more convinced of it's worth.

Also as it makes money from trends continuing meaning buying high and selling low this means data overstated can't harm the validity of the system. ie if close is higher than entry and it's made money it had to happen even if the day's high is slightly over or under the actual one.

I hope to get a site soon with orders shown on it will be easier than all this posting!

I love the idea of the high being all the stocks hitting there highs at once, it's typical american, they think we are backward but how mad is that! It's over stating how well/bad things are doing in a small way isn't it.

Stephen McCreedy