Dow 96 point entry

[bonsai]

I think today was a good example that you need a deep stop.

within about an hour of going long, you would have been down about 40pts.

I had a fib at 9830 and think I would have chickened out there
without such a stop and if it had gone to 9820 then definitely !

 

[Bill]

jpwone

Err sorry to butt in. I've clearly missed something can you help. I downloaded the spreadsheet and wondered what would happen if I dropped the sell & buy trigger below 96. Looking at the End Capital the-lower-the-number the-more-the-profit until at zero we are well on our way to clearing the national debt. I think that equates to: If the price drops below yesterday's close we sell and above yesterdays close we buy?
I've obviously not understood something here please can you help?

Bill

 

[snip]

Bill

You should understand, your face looks familiar.. are you the authors twin brother..

Cheers Snip

 

[Bill]

OOh sorry Snip & jpwone. Any similarity is purely accidental.

 

[jpwone]

LOL

Hi Bill

If you keep on dropping the entry level and the stop level you will get more trades. If you drop it far enough you will nearly always get a buy and a sell trade on the same day. One will be stopped out at the reduced stop level and the other will run to the end of the day. Thats how you end up with increasing returns if you reduce the entry level.

The spreadsheet does not capture all the intraday swings just the high and low. So, you could change the entry to plus or minus 20 points with a 20 point stop. In real use this would generate a lot of trades and a lot of stops. Slippage on this number of trades would also become a major factor.

The spreadsheet, as it stands, captures a maximum of two trades a day. One long and one short.

It should also be noted that there are plenty of days when the futures are up/down 50 or 60 points before the open and the price is skewed by this amount. If you go for too low an entry value you will not get your entries as the price will have already passed your entry level. The 96 point entry was chosen because it minimised the likelihood of not getting an entry at the required price.

Having said all this I will probably reduce my own entry level to 85 points or similar as this will probably still give good entries but will negate the effect of any slippage in both entry and stoploss (if used).

John

 

[Bill]

Ah Ok bruv I see, the spreadsheet does not take into account stops being hit. Clearly at zero the stops must be hit immediately. I noticed Sidinuk posted a modified version which apparently showed better performance. Do you think this could be mainly because the entry and exit values were lower? As you say 85 improves the outcome.

Great piece of work btw. Vote added in case you collect those things.

Bill

 

[delboy trotter]

Congrats to all.

It's very refreshing to see a collaboration of minds, working together to produce a possible strategy, that is freely available to all members of T2W.

Well done!

One small quibble guys.

GET YOUR OWN AVATARS!!! (Cheek!)

dbt

 

[snip]

Bloody triplets....

 

[daps]

Hi Jpwone

good to hear people are trying to optimise the system. so have you started using the 85 pts entry.? have you also back tested this or is 85 based on recent trading ranges.?

thanks

Daps

 

[jpwone]

Hi Daps

on recent trading ranges you would be looking at entries using a 63 point entry trigger (100 trading day average). However, the average over the longer term is still 95 points despite the recent tighter ranges. I will do some tests using a dynamic entry based on the last 100 trading days and also an entry which uses the 95 base entry and modifies this to account for the more recent average.

I suspect that the latter method of giving more weight to more recent data is the way to go.

Idea. It could be that a good rule of thumb is to use the midpoint between the long term average and the 100 day average. I will test this first.

John

 

[schoe]

Hi John, I was reading a book by Larry Williams last night called trading futures on line. In the book he uses a similar type system and when testing these systems he states ' I have tried using long term moving averages but in nearly all cases for systems the best time period is the preceding 1 to 4 days'
Don't know if you can try this on the system in between the 100 day and longer term. Maybe u could use an exponential type moving average?

Mark.

 

[jpwone]

Hi Mark

Thanks for that. There are two elements to this. First is the sample period and the number of data elements this generates and second is the overall duration of the sampling. So, an average of one day would be useless in our case as the data elements are EOD and would only give us one sample of data for our average. If our timeframe was 1 minute then a six and half hour day would generate 390 data points for our average to be calculated from. Does the book give the data sample period? 1 minute, 5 minute etc ?

John

 

[rglenn]

This is an excellent little system and there are many ways to configure and adapt it to make even more profit.

I have been using a simple system myself but could not overcome the opening spread of the spread companies.

However using Jpwone method I am able to do this.

As the market tends to be biased by 30 to 60 points by the spreadbetting companies I have used 65points as my trigger instead of 96 which is ideal for my current trading.

Many thanks again
John

 

[oarsman]

Sorry, I'm being thick here, but I don't understand how you avoid the spread on the opening bet. Yesterday's close was 9899. therefore for a bullish day today the trigger point is 9995. Using City Index on their 6 point spread (cos I'm a generous soul ) do I hit the buy when the spread is 9995/9989 or 10001/9995 or someplace else? Or as usual have I completely missed the point :cheesy:

 

[daps]

Hi jpwone

'Idea. It could be that a good rule of thumb is to use the midpoint between the long term average and the 100 day average. I will test this first.'

This sounds like a good way forward. Be good if you can keep the site updated with your findings, and importantly the current entry levels you would use.

With the 96 entry how many trades were being produced on an annual basis.? This would give a good indication regarding slippage with the sb firms. Well I guess this would be important with a varying entry also.

thanks

Daps

 

[rglenn]

Oarsman

You are not thick.

I think I have confused the issue a bit by my statement.

Just to clarify that my comments does not relate to Jpwone system.

It was only an observation which has assisted me with my system trades if I am caught at the wrong end of a 50pt move against me.

The actual spread whether 6pts or 12pts doesn't concern me.

 

[mombasa]

This looks good - only problem is that the backtesting is 1999-2001 when we had strong trends.

Not convinced if the results would be as profitable in todays market - well probably would have been great from March to Sept.

Does anybod have the 2002-2003 data they can post here?

Cheers

 

[ChowClown]

...here you go.

Sid - I'm looking at your moving average version, as a sanity check, do you make today's trades:

BUY 9944
SELL 9854

With a stop loss of 45?

Thanks.

 

[rglenn]

mombasa

I think you will find that you can use the points ranges from 56 to 86 which is inbetween 96 and you will get good results over all periods.

 

[mombasa]

Thx Chow & rglenn

will post the ss back with new data asap