Transitioning to markets available in Australia's time zone

[CanOz]

Hey CanOz, could you disclose what markets/exchange data you get from eSignal?

At the moment I'm just getting fx through esignal....it's $125 USD per month. Not really happy at the moment either, as it has some strange anomalies in some of the pairs

 

[Wysiwyg]

I'm talking about being in a flow state (described by Csikszentmihalyi and others). I've had three striking moments in my career - twice with trading and once with a sports betting account. The flow state would appear as a result of total letting go, and the result was absolute certainty about what was going to happen next (hard to explain unless you've experienced it). The certainty that flooded in left me in no doubt about how things work. Trading is 100% a mind game.

I know what you mean. The way I can explain it is this. The processing of data [input] is more exact as evidenced in the results. Decision making is not wishy-washy. One can announce (to themselves so other organisms don't sabotage) the outcome before it happens. Like being in form with the cricket bat. The ball looks like a watermelon.

 

[Wysiwyg]

At the moment I'm just getting fx through esignal....it's $125 USD per month. Not really happy at the moment either, as it has some strange anomalies in some of the pairs

Kay. Thanks.

 

[CanOz]

Only one trade got triggered last night and it got stopped out almost immediately. I think in this case i put my stop much too close, so not sure what i was thinking there but i'll chalk it up to experience, Especially with a more volatile pair like the EUR/NOK. The frustrating part is that the pair has pulled back from the highs....although this seemed like a solid level the trade was counter trend.

 

[CanOz]

I'm wanting to add a play to take advantage of the London open. Here's my thinking:

The asian session, in some pairs is very light in volume and very tight in the range. When London opens the stops accumulated on either side of the 'balance' or ''range' get flushed out and away we go.

So, to test if there is an edge here, i want to use my Open Range Breakout strategy i had coded a long time ago. I set the opening range to the Asian session, then test it using a 1:1 range to target ratio. I'll then test it on non-Asian pairs, and Asian pairs, looking mostly for win rate.

 

[Gringotts Bank]

So, to test if there is an edge here, i want to use my Open Range Breakout strategy i had coded a long time ago.

If you have the coded strategy, easier to backtest it, surely? Trade-by-trade testing would take years to even estimate if you have an edge.

 

[CanOz]

If you have the coded strategy, easier to backtest it, surely? Trade-by-trade testing would take years to even estimate if you have an edge.

Yeah, that's exactly what i said i wanted to do....just testing some pairs and tabling the results.

 

[Gringotts Bank]

Yeah, that's exactly what i said i wanted to do....just testing some pairs and tabling the results.

The degree of fiddling around with peripheral stuff which has no bearing on anything has hit a new high. You know full well that "tabling results" will not yield anything useful.

Either:

1) buy or develop a backtestable system which you can start trading OR...
2) develop your confidence to the point of being able to trade using discretion.

If your chosen method doesn't return you an income after 1 month, then you've bought/developed a poor system and need to try again. Or you listened to the wrong person when it comes to developing trading confidence, and need to try again.

 

[CanOz]

The degree of fiddling around with peripheral stuff which has no bearing on anything has hit a new high. You know full well that "tabling results" will not yield anything useful.

Either:

1) buy or develop a backtestable system which you can start trading OR...
2) develop your confidence to the point of being able to trade using discretion.

If your chosen method doesn't return you an income after 1 month, then you've bought/developed a poor system and need to try again. Or you listened to the wrong person when it comes to developing trading confidence, and need to try again.

You're really clogging up my thread with your opinions, i don't find them helpful. Really annoying actually.

I would prefer to see if there is greater than 50/50 probability of the markets doing something giving a set of circumstances before i develop a trading idea. If i already have a statistical edge, then i want to know how my desecration can may be able to improve that.

I've already mentioned that this is a new play to take advantage of the London open and the resulting surge in volume and volatility. I'm still trading my other plays, but i have time to fiddle and apparently to reply to your posts.....

 

[Gringotts Bank]

You're really clogging up my thread with your opinions, i don't find them helpful. Really annoying actually.

I would prefer to see if there is greater than 50/50 probability of the markets doing something giving a set of circumstances before i develop a trading idea. If i already have a statistical edge, then i want to know how my desecration can may be able to improve that.

Desecration... that's quite the Freudian slip.

A 50/50 probability of markets doing something...? You want to do this as a trade-by-trade manual process? You understand that takes many months? And then what... another few months of something else? A 50-50 win rate, even if you achieve it, means nothing! You have AB, but you're not prepared to run some simple backtests to save you years of manual testing? Doesn't that strike you as odd?

My help is so annoying because I'm being realistic. There's not a single trader in here who would confirm that your current approach is going to lead anywhere useful. You have to give away all the pretty charts, the ridiculous level of "preparation" and get honest with yourself.

Now if my calculations are correct, that should be enough for you to put me on ignore. Too real.

 

[CanOz]

Desecration... that's quite the Freudian slip.

A 50/50 probability of markets doing something...? You want to do this as a trade-by-trade manual process? You understand that takes many months? And then what... another few months of something else? A 50-50 win rate, even if you achieve it, means nothing! You have AB, but you're not prepared to run some simple backtests to save you years of manual testing? Doesn't that strike you as odd?

My help is so annoying because I'm being realistic. There's not a single trader in here who would confirm that your current approach is going to lead anywhere useful. You have to give away all the pretty charts, the ridiculous level of "preparation" and get honest with yourself.

Now if my calculations are correct, that should be enough for you to put me on ignore. Too real.

Desecration... that's quite the Freudian slip.

Literally a fat finger on the spell check mate....

I already have an algorithm coded mate. I have been running back tests all morning with different filter settings, different sessions and different hold times..... Why are you thinking that i'm testing manually?

 

[CanOz]

Funny how i said this:

So, to test if there is an edge here, i want to use my Open Range Breakout strategy i had coded a long time ago.

And yet you still assume this:

The degree of fiddling around with peripheral stuff which has no bearing on anything has hit a new high. You know full well that "tabling results" will not yield anything useful.

You want to do this as a trade-by-trade manual process? You understand that takes many months? And then what... another few months of something else? A 50-50 win rate, even if you achieve it, means nothing! You have AB, but you're not prepared to run some simple backtests to save you years of manual testing? Doesn't that strike you as odd?

How on earth did you put 1+1 and get 3?

 

[CanOz]

These are the test results on three pairs.

First i tested the strategy of taking a position on the break of the Asian session range, putting a stop at 1 x the range and a take profit at 1x the range 1:1 R/R

I also tested exiting at the close and exiting at the target. I also have a narrow range day filter as a condition that tested as well. So looking for a narrow ranging day to setup on.

Then i tested the same strategy but this time i combined both the Asian and the London ranges. Then tested to the close and to the target, filter on and filter off.

The GBPUSD tends to be much more likely to trade to its target after taking out the range, regardless of the filter.

The EURUSD is more likely to trade to its target after taking out the combined range as well, filter doesn't have any improvement.

The EURGBP is still pretty much a coin toss.

An edge can be defined as something that occurs more than random. We are only taking profit at 1 x risk. The GBPUSD with the combined range strategy may have potential to develop into a full strategy.

 

[Gringotts Bank]

You used the word "tabling results" rather than "backtesting". Alright good. So you want to optimize with a optimization target of 'win%', yes? That's good too, but you need to make sure the win rate holds up with walk forward testing because high win rates are easy to generate and easy to curve fit to data. Profitability should be at least 'break even' if you're going to use this approach. Does it pass both these tests?

If it does, what methods are you going to use to optimize your discretionary decision making so that you choose more of the right trades than the wrong ones?

For someone who trades a combined system+discretion, the less profitable the backtest, the better your discretion has to be. Also, try to avoid getting a high win% at the expense of having a few massive losers, unless your discretion is super tuned in to the market. One slip up and you can do massive damage.

 

[CanOz]

You used the word "tabling results" rather than "backtesting". Alright good. So you want to optimize with a optimization target of 'win%', yes? That's good too, but you need to make sure the win rate holds up with walk forward testing because high win rates are easy to generate and easy to curve fit to data. Profitability should be at least 'break even' if you're going to use this approach. Does it pass both these tests?

If it does, what methods are you going to use to optimize your discretionary decision making so that you choose more of the right trades than the wrong ones?

After all that, i don't even get an apology?

I plan on do some manual statistics on the successful trades, since there are not many of them. I'm going to look for a set of conditions that will involve hourly volatility and daily range of the setup session and the session before.

I can also do some (Back)tests on the most successful pair by tightening up the stop and going for larget targets. Obviously the win rate will drop, but the wins should get larger. This would indicate that i may be able to use a trailing stop.

Would you like to suggest anything else?

 

[CanOz]

I have tested the GBPUSD pair with a target of three times the range and a stop of 1x the range -20 ticks (inside the range). The win rate has dropped to 48%, but the average win to average loser is 3.21 times now.

This is a profitable strategy in its own right. There is a caveat or two though. The risk is the average hold time is now longer than 5 days, so its holding over the weekend in many cases. Also, there are no commissions, but ample profit to cover. There are only 25 trades for a whole year, yet it is not being filtered, the system is not taking new trades while in a trade.

 

[CanOz]

The results i'm getting on other pairs are very encouraging. Even without looking at any other conditions, all the equity curves are very positive. Some of the trends that it has caught are so long that it only has taken a few trades a year. I guess this is in part to having only an initial stop and a 3 x take profit. This is the auto trader dilemma, stops hurt performance, what kind of a trailing stop can one use to capture eve ngreater reward without getting stopped out too early?

I'm going to run a portfolio level back test. Then perhaps some small portfolio backtests but using OOS data.

I think i may have crashed NT7....

 

[Gringotts Bank]

After all that, i don't even get an apology?

I plan on do some manual statistics on the successful trades, since there are not many of them. I'm going to look for a set of conditions that will involve hourly volatility and daily range of the setup session and the session before.

I can also do some (Back)tests on the most successful pair by tightening up the stop and going for larget targets. Obviously the win rate will drop, but the wins should get larger. This would indicate that i may be able to use a trailing stop.

Would you like to suggest anything else?

1) not many trades + 2) optimization = curve fitting. How many years of data are you backtesting on with an hourly time frame?

Curve fitting is ok, so long as it walks forward on unseen data, otherwise it's not a real edge.

Since you are overlaying a discretionary component, what method are you using to maximize the quality of your decision making?

 

[ThingyMajiggy]

1) not many trades + 2) optimization = curve fitting. How many years of data are you backtesting on with an hourly time frame?

Curve fitting is ok, so long as it walks forward on unseen data, otherwise it's not a real edge.

Since you are overlaying a discretionary component, what method are you using to maximize the quality of your decision making?

LOL I would say shut up, but to be honest I'm enjoying it too much, it's a cross between a face palm and this...



All the best Can, dunno how you put up with it

 

[ThingyMajiggy]

1) not many trades + 2) optimization = curve fitting. How many years of data are you backtesting on with an hourly time frame?

Curve fitting is ok, so long as it walks forward on unseen data, otherwise it's not a real edge.

Since you are overlaying a discretionary component, what method are you using to maximize the quality of your decision making?

Clean out your inbox, I'm trying to reply to your attempt at abusive PM's Nice job on calling me a c-bomb before it got deleted too, classy