Equity curves - how smooth do you like yours to be B4 you consider trading it?

[JTrader]

Equity curves - how do you like yours to look B4 you trade those system setttings?

Hi

if you backtest a mechanical system over a given period of time with different settings/parameters, you may end up with a profit of equal amounts, but the equity curve may tell a different story.

EG.
Settings 1 - may show a smooth ascending equity curve.
Settings 2 - may show a a non-smooth equity curve with peaks & troughs b4 a steep ascent towrds the end of the time period.
settings 3 - may be flatlining until the middle of the equity curve b4 a steep ascent.
settings 4 may have performed superbly in the 1st two thirds of the test by ascending smoothly, b4 tailing off into a descending equity curve in the last 1/3 of the time period tested.

etc. etc. - there are many other possible variations.....

To me, settings 1 is the most consistent performer, despite the fact that all 4 settings made the same profit come the end of the time period tested. Settings 4 would probably concern me the most, as most recently these settings have been losing.

How important is/should be the appearance of the equity curve, in your opinion, in influencing your decision to trade those system parameters live or not.

Thanks.

 

[the blades]

How important is/should be the appearance of the equity curve, in your opinion, in influencing your decision to trade those system parameters live or not.

Thanks.

A smooth equity curve is the most important factor for me - whatever the system stats declare.

I also want to see the slope maintained through significant changes in the market's direction, especially if the system trades one side only.

UTB

 

[JTrader]

A smooth equity curve is the most important factor for me - whatever the system stats declare.
UTB

Yes, for me, it certainly seems a lot more re-assuring when the EC is smooth & reasonably diagonal in ascending slope towards 45 degrees (ideally!)- i.e. consistent/steady etc. I'm starting to look at the EC's more closely.

 

[Hoggums]

Hmm - 45 is easy with the right scale on the graph! Obviously the smoother the better - but usually smoother means less risk -> therefore less profit - so you have to look for a happy medium. I'm quite prepared to accept a 20% total DD so I tend to scale my systems for that using worst case scenario data I can find.

 

[the blades]

Hmm - 45 is easy with the right scale on the graph! Obviously the smoother the better - but usually smoother means less risk -> therefore less profit - so you have to look for a happy medium. I'm quite prepared to accept a 20% total DD so I tend to scale my systems for that using worst case scenario data I can find.

Interesting.

System 1 = 15% annual return and 5%DD
System 2 = 30% annual return and 15% DD

Which would you trade?

UTB

 

[mr_cassandra]

The home page says the following today: Adaptation to Market RealitiesBy Joe Ross There is a need to adapt and change your behaviour relative to the market because the markets are ever-changing. Mechanical systems may be workable, but for only a short time so you must learn to trade what you see on a chart.

any thoughts on the part in bold ?

 

[the blades]

The home page says the following today: Adaptation to Market RealitiesBy Joe Ross There is a need to adapt and change your behaviour relative to the market because the markets are ever-changing. Mechanical systems may be workable, but for only a short time so you must learn to trade what you see on a chart.

any thoughts on the part in bold ?

Probably true - but if you trade systems with a smooth equity curve it's then easier to define when something has changed.

I currently operate about 20 systems - no doubt these will change / evolve in future.

UTB

 

[JTrader]

The home page says the following today: Adaptation to Market RealitiesBy Joe Ross There is a need to adapt and change your behaviour relative to the market because the markets are ever-changing. Mechanical systems may be workable, but for only a short time so you must learn to trade what you see on a chart.

any thoughts on the part in bold ?

He doesn't qualify this statement in anyway. How short is a short time etc. does he mean b4 total abandonment or minor modification etc.....
I have never blindly followed anyones advice, let alone the advice of an author, who is short on details (not that i'm saying JR is, as i dont know him from Adam).....There's no reason why MS's shouldn't work over the loing haul.

 

[JTrader]

Hmm - 45 is easy with the right scale on the graph! Obviously the smoother the better - but usually smoother means less risk -> therefore less profit - so you have to look for a happy medium. I'm quite prepared to accept a 20% total DD so I tend to scale my systems for that using worst case scenario data I can find.

True. the appearance of the EC to some extent depends by how much you compact the diagonal line.
Balance is always key to most things.
While consistency is preferable (45' ascending EC -without any EC scale manipulation), if the EC showed a flat period B4 the curve began ascending, because the system wasn't losing money, this is perhaps acceptable also.

 

[mr_cassandra]

I look at drawdown in three ways.

1. drawdown 'during' a trade
2. drawdown from a completed trade
3. drawdown from worst string of losing trades

Which would you say is most meaningful ?

 

[JTrader]

I look at drawdown in three ways.

1. drawdown 'during' a trade
2. drawdown from a completed trade
3. drawdown from worst string of losing trades

Which would you say is most meaningful ?

I don't fully understand what 2 means

Cheers.

 

[mr_cassandra]

I don't fully understand what 2 means

Cheers.

2. would be the amount your account was drawdown by the completed losing trade.

What I'm driving at is what can happen 'during' a trade that ends up fine versus a completed losing trade.

 

[JTrader]

2. would be the amount your account was drawdown by the completed losing trade.

What I'm driving at is what can happen 'during' a trade that ends up fine versus a completed losing trade.

So like a losing trade that didn't hit you maximum SL, but maybe got 60% of the way there?

 

[mr_cassandra]

So like a losing trade that didn't hit you maximum SL, but maybe got 60% of the way there?

let me give you a real life example. My own work was originally designed to target the major seasonal lows in the US markets, usually but not always in the fall.

It goes without saying they can't be targeted perfectly so my goal was to get as close as possible without missing any.

With the above in mind, during rough declines like the collapse of LTCM years ag, the trade can see a substantial drawdown 'during' the trade, yet it completes at a profit.

My completed trades worst case drawdowns was from 3 losers in a row, piling up a loss of 8%.

But during trades that ended with a win, there have been worst case drawdowns of 34%, and 26% though admittedly very infrequent.

Those numbers are all based on using 2X proshares etf's with occasional margin.

Sorry for rambling but what would you take as more meaningful, the 34% 'during' a trade which ended at a profit or the 8% from 3 finished losers in a row, as regards the subject & meaning of drawdown?

 

[the blades]

let me give you a real life example. My own work was originally designed to target the major seasonal lows in the US markets, usually but not always in the fall.

It goes without saying they can't be targeted perfectly so my goal was to get as close as possible without missing any.

With the above in mind, during rough declines like the collapse of LTCM years ag, the trade can see a substantial drawdown 'during' the trade, yet it completes at a profit.

My completed trades worst case drawdowns was from 3 losers in a row, piling up a loss of 8%.

But during trades that ended with a win, there have been worst case drawdowns of 34%, and 26% though admittedly very infrequent.

Those numbers are all based on using 2X proshares etf's with occasional margin.

Sorry for rambling but what would you take as more meaningful, the 34% 'during' a trade which ended at a profit or the 8% from 3 finished losers in a row, as regards the subject & meaning of drawdown?

Steve - it would be the 34% for me - though it would have to be in context with the overall gain - ie through the equity curve or a sharpe / sortino ratio.

Some people think the money's not lost intil the deal is cloased. Not for me.

UTB

 

[JTrader]

let me give you a real life example. My own work was originally designed to target the major seasonal lows in the US markets, usually but not always in the fall.

It goes without saying they can't be targeted perfectly so my goal was to get as close as possible without missing any.

With the above in mind, during rough declines like the collapse of LTCM years ag, the trade can see a substantial drawdown 'during' the trade, yet it completes at a profit.

My completed trades worst case drawdowns was from 3 losers in a row, piling up a loss of 8%.

But during trades that ended with a win, there have been worst case drawdowns of 34%, and 26% though admittedly very infrequent.

Those numbers are all based on using 2X proshares etf's with occasional margin.

Sorry for rambling but what would you take as more meaningful, the 34% 'during' a trade which ended at a profit or the 8% from 3 finished losers in a row, as regards the subject & meaning of drawdown?

Well for me, and my way of trading, a 34% or 26% drawdown in one trade, or even 10 trades is just not an option. I would have an SL in place, which is typically based on a % of my capital, and so this would take me out b4 hand.

I consider the closed trades results the most significant factor. Though i also like to know/monitor how close price got to my SL during the trade, but if the SL was not hit, i would not lose too much sleep over it. But ongoing monitoring of the SL size is worthwhile IMO.

 

[mr_cassandra]

Well for me, and my way of trading, a 34% or 26% drawdown in one trade, or even 10 trades is just not an option. I would have an SL in place, which is typically based on a % of my capital, and so this would take me out b4 hand.

I consider the closed trades results the most significant factor. Though i also like to know/monitor how close price got to my SL during the trade, but if the SL was not hit, i would not lose too much sleep over it. But ongoing monitoring of the SL size is worthwhile IMO.

Understood and thanx for your input, as well as theblades. I have often said to others its one thing for me to sit here with all the program data and years at it and say 'stay the course' but I can totally understand anyone else feeling darned nervous. Even during the august subprime swoon this year, the draw was 20% before going on to end with a profit. That one occurred with everyone in my yahoo group watching.

The tough part is I have made code run after code run trying for a workable get out at X theen re-enter at Y and every single one so far cost you more profit than standing pat.

I'd love to solve it though as it would raise their comfort level immensely.

 

[TickCOM]

3.

Max drawdown is supposed to show the max loss you may experience when you start trading the algorithm.

You can be very unlucky starting with the new algorithm just as it starts a period of losing, consider it to be a worst case indication.

 

[mr_cassandra]

max loss question

You mean max loss from that algorythms closed trade?

How about max draw during an uncompleted trade? (even if it ends up with a profit if held until completion)

Would there be any way to put numbers to this on a sliding scale, IE if the system was generating X% annually I could tolerate an intratrade draw of y% and a realized loss of z%

3.

Max drawdown is supposed to show the max loss you may experience when you start trading the algorithm.

You can be very unlucky starting with the new algorithm just as it starts a period of losing, consider it to be a worst case indication.

 

[TickCOM]

You mean max loss from that algorythms closed trade?

How about max draw during an uncompleted trade? (even if it ends up with a profit if held until completion)

Would there be any way to put numbers to this on a sliding scale, IE if the system was generating X% annually I could tolerate an intratrade draw of y% and a realized loss of z%

How about max draw during an uncompleted trade?

Thats a very good question. If you include max drawdown during the trade, this will be to make sure you wont get below the margin required by your broker. If not included, it will indicate the risk profile of the algorithm it self.

IE if the system was generating X% annually I could tolerate an intratrade draw of y%

You are talking about anually - please note that you can not base your automated trading of algorithms trading a few times a week:

1) This timeframe is too short. The way prices are moving on the stock exchange are changing, an algorithm which worked last year may not work this year.

2) Its close to impossible to backtest such algorithm, since you need to backtest on 3-8 years of historical data, thus on historical data which may not work any more.

3) Before you trade for real, you have to trade simulated using live data feed, but on a paperaccount. Thus you have to test simulated during a year - which is not operational.

4) Take care of public algorithms, they are posted since they didnt work for the developer. Especially be alerted on bakctests using daily bars, trading a few times a week.