True symmetry not currently achievable by default...Must use custom template to get it? Please give us an option so we can have 100% symmetry.

Voted for this task.

I am struggling to understand exactly what you are talking about. If you want to make a strategy that only trades long and make a separate strategy that only trades short, you can do that. Symmetry only refers to opposite trading rules which are generated automatically. So what do you really mean by "true symmetry" when referring to trades that cancel each other out? Are you saying you want a trade to be entered in the opposite direction when the other trade takes profit?

I want longs and shorts together in one strategy and the algo to always be symmetrical. Occasionally, a strategy will be produced where sometimes long signals fire by themselves, and sometimes short signals fire by themselves but sometimes the long and short signal are firing simultaneously. I'm saying that in the case where both signals fire simultaneously I don't want just longs to be placed by themselves. That's not  proper symmetry.

If both long and short signals are true then hedge would be symmetrical.
If both long and short signals are true then do not place any orders at all would be symmetrical.
If both long and short signals are true then place longs but do not place shorts is NOT SYMMETRICAL.

If long and short signals are both firing at the same time, it means both conditions are true for the trades to execute. This may be a bug in how the strategy is made. I have not seen a strategy where symmetrical rules are both true at the same time as the point is for the rules to be of opposite conditions. If they are firing simultaneously by accident or you are somehow building them in a way to do this on purpose, I would recommend citing specific examples where they are supposed to both fire off at the same time and only one is instead. 

There is no way to even know if they ARE firing simultaneously because only a long will be placed by default.

But off the top of my head here are a couple signals which have potential to sometimes overlap buys and sells even when properly symmetrical:
Bear power / Bull Power 
ADXdi- / ADXdi+ 

Also good old buysignal=always true sellsignal=always true. We may want to place limit orders for both signals / price levels simultaneously for instance.

Also close above lower band / close bellow upper band.
Close above low ANYTHING  / close below high ANYTHING.

Bar closes bellow SMA + x * ATR / Bar opens above MA - x * ATR

Lowest is rising for x bars / Highest is falling for x bars

Price above a fibo level  / Price bellow the opposite fibo level

"#9

bentra 13 minutes ago

Bar closes bellow SMA + x * ATR / Bar opens above MA - x * ATR"

oops that one was garbled lets try
Price bellow Moving average + x / Price above Moving average - x

Some High > Other High / Some Low < Other Low

About 5% of all strategies from random building are going to produce at least some simultaneous long and short signals. Way more than 5% if you include true/true strategies.

As for SQX teams suggestion to use a custom template: As far as I can tell, using a custom template to get true symmetry is somewhat limiting the building process. We couldn't for instance have it randomly choose how many conditions, we'd have to have a static number of conditions.

So yeah we really need this bug fixed ahem I mean "feature."

I guess the way you're phrasing this is the confusing part. Are you saying more complex strategies are not producing symmetry when this setting is turned on? Can you provide sqx examples where the rules are not symmetric? 

Or are you saying SQX lacks the ability to create these type of rules in the first place? I would understand this line of reasoning more. I just didn't understand the part where you were talking about simultaneous orders and a long bias as you can direct SQX not to take long trades at all.

As for the definition of symmetry: Symmetry is not exclusive to longs and shorts in isolation of each other. (Consider a symmetrical art piece. There can be a center area which takes characteristics of both the left and the right side. Like this... <<<<><><>>>>) So, technically, it would be incorrect to consider a strategy only symmetrical if it fires longs in isolation of shorts. I also think there are many reasons why we should not expect this of SQX regardless of what name we give it, some strategies SHOULD at least sometimes fire and place longs and short simultaneously and the exits should be allowed to play out. You disagree? You think any algo should only fire one signal at a time? What about a simple, longs and shorts always true strategy that produces a stop buy order at the highest high level and a stop sell order at the lowest low? This could be a beautiful symmetrical breakout strategy but it is being broken by the  "and Not LongEntrySignal " applied only to short conditions!!!

I'm assuming we already know SQX is giving random strategies by design and will contain some normal looking strategies which will fire sometimes long and short together. You're saying this is not true? 

Any bears power bulls power based strategy will eventually fire both signals simultaneously. Since one is based on high's and the other is based on low's. If lows are increasing and high's are simultaneously decreasing then bears power and bulls power indicators move in a way which eventually will fire a long signal and short signal simultaneously. Just do the math.This may be a rare occurrence depending on the variables but it should be an easy one for people to wrap their head around, the calculation for this indicator is rather simple and can not be easily fixed to give only a long or a short in a symmetrical way and clearly these strategies are not all bugs either.

The generator does produce strategies such as "If close above low" and the reverse is "If close below high." Again, I'm assuming this is by design because mathematically, symmetry is not exclusive to two conditions in isolation of each other. And since we already know some indicators can produce long and short simultaneously why would we consider this one in particular to be a bug?

And what about long_signal and short_signal always true strategies???  "and Not LongEntrySignal " applied only to short conditions breaks symmetry by causing it to never take shorts and always longs.

I'm sorry but I cannot agree with what you are saying and I have read this many times to make sure I understand your point. Symmetrical rules only work by having opposite rules for longs and shorts. If a long strategy consists of trading when above a moving average, and the corresponding short strategy consists of trading when below that same moving average, both conditions will not be true at the same time and thus trading in both directions cannot occur. The whole point of using symmetrical rules is to find a method that profits for upswings in the market as well as downswings. Both a downswing and an upswing cannot occur at the same time so simultaneous trading should never happen.

Furthermore, if you were to enter a trade at the exact same time on both ends, all you would be doing is selling at the bid and buying at the ask price. No matter what happens in this scenario, you will pay the spread.  By definition, this cannot be profitable because the buy position would close out at a negative difference of the spread when the sell position is closed out. The only way to profit this way is if you were able to trade with the bid and ask invertedly priced. This is impossible with directional algorithms. You would need to be doing something like arbitrage and even then, that's not something SQX can do as this does not rely on any market data do find these type of trades. What you are describing will result in a consistently losing algorithm just because of the spread alone.

"The whole point of using symmetrical rules is to find a method that profits for upswings in the market as well as downswings. Both a downswing and an upswing cannot occur at the same time so simultaneous trading should never happen."

But sometimes the indicator or logic fires both signals at the same time (I have provided many examples) such as the bears power bulls power example or when sell_signal=true always and buy_signal=true always... What to do then? 

"Furthermore, if you were to enter a trade at the exact same time on both ends, all you would be doing is selling at the bid and buying at the ask price. No matter what happens in this scenario, you will pay the spread.  "

True and not a bad argument to have MARKET ORDERS simultaneous signals cancel each other out to retain symmetry. I'd debate that I'd like that to be optional though. I'd like to see some tests where the exits play out from a hedge. But yes it can be profitable to begin a trade from a hedge depending on the exits and what the market does. Consider a hedge with both long and short trades and they both have a tight SL and a far away TP. It's not impossible if the long side gets to TP and the short side gets stopped out that overall the strategy would still make money. Although such a strategy could be made more optimally using the SL points as the entry points avoiding the hedge entirely.... it could still be profitable to start with a hedge, we pay more commission and spread but could still be profitable. 

All of your points only apply to MARKET ORDERS and this bug is effecting LIMIT/STOP ORDERS TOO. My example about simultaneous buy and sell signals always true with a buy stop at the high and a sell stop at the low is a solid example of a potentially profitable symmetrical STOP ORDER strategy which would be broken by this bug. This is also a good example of a strategy with simultaneous longs and short signals firing and longs and shorts limit order being place without entering market in both directions necessarily. 

Here's an example that came from builder. It SHOULD be symmetrical (the code I show is symmetrical) and it SHOULD place both long and short stop orders well above and bellow the close. But it is broken by the bug and only places long orders. This is only one small piece of the puzzle though that is retention of symmetry.

Here's a MARKET order strategy example from builder of some logic which will fire simultaneously often long and short signals together... what are we going to do about it? Right now it discards shorts and places longs when both signals are true. I'm saying that is not a symmetrical solution. I'm saying we should either hedge or not place any orders when both signals fire.

Here's an example of a market order  strategy from builder which fires longs and short simultaneously but not as often. When it does fire them simultaneously, only the longs are taken and the shorts are discarded:

Here's a rather extreme bulls/bears  market order example of an indicator strategy from builder where the trading signals are both symmetrical and firing longs and shorts simultaneously fairly often. When they are both true, only longs are taken. I'm saying this is not a symmetrical solution. It is not fair to currency traders or traders who want true symmetry.

"Both a downswing and an upswing cannot occur at the same time so simultaneous trading should never happen"

Yes I think that's a fair argument if we are talking about MARKET entry strategies. For stop and limit entries I think it's ok to have longs and shorts placed simultaneously (if they are far away from each other) even if you don't want to see hedging strats as it will not necessarily result in a hedging strategy to do so, part of the system is the market must give further confirmation of a direction by moving towards and triggering the stop or limit order. But anyways, back to market strats... what will we do about the bulls/bears example? It is symmetrical but still it fires sometimes simultaneous long and short signals fairly often. It is working as intended. It is not a bug. What will we do? It is not symmetrical to simply discard the shorts and only place the longs. It is not practical to revamp this and every other code to guarantee longs and shorts are isolated from each other while retaining symmetry, especially the two-part indicators like bears/bulls or DI- and DI+. This is why I say that it might be best to have the simultaneous orders cancel each other out in a MARKET order situation. It's quite an easy and simple fix to retain symmetry to just cancel MARKET orders for both longs and shorts when we inevitably run in to simultaneous longs and shorts signals. This feature request is asking for this as an option to retain symmetry (without having to limit some of the builder functions via custom template.)

 A hedge with market orders simultaneously is useless as it will always be more effective to find an edge in a strategy in one direction and then find a strategy that has an edge in the opposite direction once the first trade has completed in order to prevent unnecessary drawdown. So no matter how you cut it, market orders done simultaneously on the same asset will not be a viable solution.

However, there may be a benefit in placing two limit orders X distance away from each other and canceling the other one when the first order fills. This would be similar to an idea I submitted earlier where instead of canceling the other limit order, it was used as a stop-loss: https://roadmap.strategyquant.com/tasks/sq4_1859. This is essentially the same idea with the understanding that the second order would be canceled after the first trade is entered. Or you could even cancel the second limit order after the market moves x pips in your direction to have a mixture of both. 

The only way to hedge and use market orders is to do this between different markets. For example, finding some strategy set of rules where you can buy EURUSD and sell AUDUSD and make a profit after x amount of time or when x exit condition occurs. I have also submitted a request for that here: https://roadmap.strategyquant.com/tasks/sq4_2829.

I wouldn't necessarily consider hedging "symmetry" as the rules to enter aren't based on any symmetrical indicator block rules between long and short. I would also argue that any blocks that cause simultaneous entry currently would stand to be a bug and instead it should not be indicator blocks that determine simultaneous entry but be a specific trading mode just as fuzzy logic is a specific trading mode. Using indicator rules to define simultaneous entry is contradictory by nature and hence my confusion with the way you laid out your feature request. 

Also, in the examples you provided, it seems to me that firing orders simultaneously is a bug. I see no reason why this should happen. Curious to get Mark's input.

"A hedge with market orders simultaneously is useless as it will always be more effective to find an edge in a strategy in one direction and then find a strategy that has an edge in the opposite direction once the first trade has completed in order to prevent unnecessary drawdown. So no matter how you cut it, market orders done simultaneously on the same asset will not be a viable solution."

I do agree that a hedging position is never truly optimal for performance.

"The only way to hedge and use market orders is to do this between different markets. For example, finding some strategy set of rules where you can buy EURUSD and sell AUDUSD and make a profit after x amount of time or when x exit condition occurs. I have also submitted a request for that here: https://roadmap.strategyquant.com/tasks/sq4_2829."

 Short EURUSD and Long EURUSD together is considered a "hedge" no matter how you entered. The hedging solution is a little weird. We don't need it for symmetry, but I think it could be interesting. There are other symmetric solutions to solve this issue besides hedging.  Sometimes my breakout strat winds up hedged as price fakes out quickly from one way to another and triggers both stop entry orders (where SL is a bit wider) I admit that I could code it in a way that avoids the hedged positions. (Instead of going hedged we can go flat and then just re-enter at the point where one side of the hedge stops out...) and get slightly better results (less spread and commission.) But it's more complicated to code and track orders that way, it happens so rarely so I just let the hedge play out and the sides individually exit according to their own rules. 

Don't you have a couple of different strategies that sometimes wind up one being long and one being short in the same market? The market isn't going to go both ways and we both agree that hedging isn't optimal for performance but yet many people are still sometimes long and short simultaneously.... food for thought. The long and short of a single EA can be thought of as two separate strategies and a hedge can be allowed to play out with individual exits and you could still have a winner. It's not optimal but it's not necessarily terrible and it is easier to code and track orders (at least in MT4)

"However, there may be a benefit in placing two limit orders X distance away from each other and canceling the other one when the first order fills. This would be similar to an idea I submitted earlier where instead of canceling the other limit order, it was used as a stop-loss: https://roadmap.strategyquant.com/tasks/sq4_1859. This is essentially the same idea with the understanding that the second order would be canceled after the first trade is entered. Or you could even cancel the second limit order after the market moves x pips in your direction to have a mixture of both. "

Good symmetrical ideas but more complex to code and track orders and signals. I believe the 0 condition state which just says longs and shorts always true was purposely put in to SQ a very long time ago for a reason though and is not a bug... Some people do want simultaneous signals always true. My guess is that was geared for limit and stop orders. The long and short  orders can be thought of as separate strategies and left to play out by their own rules. I think the popular opinion will be that it is ok to have symmetrical simultaneous limit or stop entries on the chart that are just left to play out as their original rules state. But I'd be happy if they resolved it your way or any way that retains symmetry because just cancelling shorts and letting longs go is clearly not symmetrical. This feature request is asking for the option to allow longs and shorts to be placed simultaneously in a STOP or LIMIT order scenario in order to retain symmetry (without having to limit some of the builder functions via custom template.) as that would be the simplest, most practical solution. An option to not place any STOP or LIMIT orders during simultaneous signals (until only one is true) might be a simple way to appease Keincs camp too in case these blocks I submitted are not declared bugs. 

"Also, in the examples you provided, it seems to me that firing orders simultaneously is a bug. I see no reason why this should happen. Curious to get Mark's input."

So 5% (or way more if you include true/true strats) of what comes out of builder is bugged? I suppose it's possible to re-code every offending indicator block to guarantee long signals and short signals do not overlap but that's not a small job and will require some ingenuity... You get what I'm saying about bears/bulls DI-/DI+ right? Occasionally they move together so there is not an easy way to avoid simultaneous signals while retaining symmetry except to either allow the hedge or to have them cancel each other out until only one is true.  I'd be happy if they resolved it your way or any way that retains symmetry because just cancelling shorts and letting longs go is clearly not symmetrical. However, It's quite an easy and simple fix to retain symmetry to just cancel out MARKET orders for both longs and shorts when we inevitably run in to simultaneous longs and shorts signals (until only one is true.) This feature request is asking for this as an option to retain symmetry (without having to limit some of the builder functions via custom template.)

  "Fact: builder produces 5% or more of strategies that have logic which results in some simultaneous long and short signals. (Some proof bellow.)
-If these are all bugs, let's fix them all in a way which retains symmetry."

I agree with this. This should essentially never happen because of symmetry should never cause short entry conditions and long entry conditions to be true at the same time. This makes no sense and will force the spread to be traded at an immediate loss on market orders and limit orders may be even worse if they are submitted further away and a spike occurs in the market. The case where I can see this happening? Say you have a Bollinger band strategy that buys when the top band goes up and sells when the bottom band goes down. This should not happen as this would cause the long and short entry to trigger at the same time, resulting in a potentially massive loss. Anywhere that indicator block conditions cause this to happen is most likely a bug and should not be expanded on as all scenarios can't benefit the strategy. 

In order to achieve the type of symmetry that you are discussing, you would need to make a hedging mode, similar to having mean reversion or fuzzy logic modes. The hedging mode would handle using simultaneous order handling with limit orders. This solves the issue of retaining the hedging logic without basing it off of indicator logic. If this is done, it should only be done with stop-limit orders and users should have the choice of how to handle the logic for those orders i.e. Removing one order after a trade is triggered/after a trade is triggered and move x ticks/after x bars. This should never be done with market orders as the spread will always be paid, causing any profit made by the exit methods to be an inconsequential edge only found between the exit methods instead of at the time of entry causing unnecessary margin and loss due to spread tied up in trading conditions. Making a new mode to handle this for limit orders only is the smartest way to go about achieving what is being discussed.

Voted for this task.

Voted for this task.

since march 2018 EU and everything else is Short BIAS market.  Most strategies fail this might be a clear cause. However when you have conflicting signal non should be taken. probably flat or chop market.

Voted for this task.

Voted for this task.

Voted for this task.

Agree with bendx77 that a true symmetric strategies will follow this rule:

 

if LongEntrySignal and Not ShortEntrySignal
// Long action – Market, Stop or Limit

if (ShortEntrySignal and Not LongEntrySignal)
// Short action – Market, Stop or Limit

 

The standard template will favour long entries making the strategy unsymmetrical .

it is not good to add more complex options to SQ. All this can be achieved with custom strategy templates

I think a checkmark option that says "default place long orders and not short orders when long and short are both true" with a descriptive info popup would not be complicated and would add some transparency. I think having a long bias building process (which only a programmer might notice the cause) without any notes to the user about that is a little misleading. Currency markets don't have a natural long bias and some traders believe in using properly symmetrical strategies.

Please vote if you agree.

Voted for this task.

Voted for this task.

Voted for this task.

Voted for this task.

Voted for this task.

Voted for this task.

Voted for this task.