Index futures automation

[Laissez Faire]

Well that's a balance chart, not equity. So I'm not averaging down trades, but i do widen my reverse (stop) and target and also adjust the position to recover (not quite martingale, it follows a sequence after few reverses fail).

But I do emphasize consistency over the large inconsistent profit targets, so the wins are capped as you say.

What's a balance chart?

Regardless, if your largest win is $1411 and your average daily gain is in the ballpark of $50-100 the logical conclusion would be that you have had a drawdown in that ballpark which you recovered from to end the day around your average daily gain.

So, in summary, you cap the wins while not capping the losses. This seems to be the opposite of letting your winners run and cutting your losses short. So far it seems to have worked out very well, but since you brought up the subject of black swans, I suppose it's possible you'll one day won't be able to recover from an intraday drawdown and would have to realize a larger loss (relatively speaking).

I wonder if there'd be a way for you to squeeze out more of your winners, i.e., maybe scale out of your positions by letting a winner run. For example booking a profit on one contract and letting the other run with a B/E entry or similar.

Regardless, I'm not giving you any advice here as it seems to be working very well for you, but your intraday drawdown figures does raise some concern for that black swan showing up one day just by the law of large numbers.

 

[hilmy83]

What's a balance chart?

Regardless, if your largest win is $1411 and your average daily gain is in the ballpark of $50-100 the logical conclusion would be that you have had a drawdown in that ballpark which you recovered from to end the day around your average daily gain.

So, in summary, you cap the wins while not capping the losses. This seems to be the opposite of letting your winners run and cutting your losses short. So far it seems to have worked out very well, but since you brought up the subject of black swans, I suppose it's possible you'll one day won't be able to recover from an intraday drawdown and would have to realize a larger loss (relatively speaking).

I wonder if there'd be a way for you to squeeze out more of your winners, i.e., maybe scale out of your positions by letting a winner run. For example booking a profit on one contract and letting the other run with a B/E entry or similar.

Regardless, I'm not giving you any advice here as it seems to be working very well for you, but your intraday drawdown figures does raise some concern for that black swan showing up one day just by the law of large numbers.

Your post does trigger a thought that I have when talking about risk in general. The way I see risk is not based on individual trades (like how much risk i need to put on a trade). I see risk as a whole from the perspective of risk of ruin (probability of blowing up).

Every strategy will have a probability of ruin. I don't care if it's a high reward to risk strategy or like mine with reverse R:R. If you are placing a trade, you're risking a portion of your capital. Whether that is 1% of your capital or 1 contract. If you lose that trade, well, you're not going to stop trading are you? You're going to continue.

Then if the next trade is a loser, you will still continue until you decide that your risk for the day/week/month is met and you stop trading OR continue trading. I'm in the camp of as long as the trading day is open, I will continue until I meet the target (minus comms)

But in either case, whether you stop early or continue trading, there is a "chance" that you can blow up your account. Example a 2 % strategy will lose 80% of the capital after about 79 consecutive losses. That is pretty low occurence. You need to kow your win/loss % and the sequence count (since we're trading futures, I take 80% loss as risk of ruin since that's the max cap futures broker will take before they auto liquidate you or stop your trading for the day).

You can lower that probability by risking small or having a max loss day like I mentioned above, but that risk of ruin is NEVER going to be zero, unless you stop trading altogether with your capital still intact.

So the way I lower my probability of ruin is trhough quick targets (amongst other things like picking the ideal time to trade, the right opening range, the right market, etc.)

I'm assuming that the combination of all those things above lead to a high probability that the strategy will survive "longer". I'm very aware that there is a chance of a blow up.

When I tested this with limited data (3 years) i didnt see the event. And now about 6 months in live, it's still functioning. But like I said in previous post, there is a price action scenario where it would kill this strategy. We'll just have to wait and see when that probabilty will hit.

 

[hilmy83]

On unrelated topic, if you are using Quantower, do not use download link that AMP provides. They advertise all features available for AMP customers, which is not true based on the link they provided.

If you use the AMP download link, you will see these features:

But if you actually just download straight off Quantower's site, you actually do get all the features. As you can see a few more features available.

I just realize that literally last week.

 

[Overnight]

On unrelated topic, if you are using Quantower, do not use download link that AMP provides. They advertise all features available for AMP customers, which is not true based on the link they provided.

If you use the AMP download link, you will see these features:

View attachment 321918

But if you actually just download straight off Quantower's site, you actually do get all the features. As you can see a few more features available.

I just realize that literally last week.

View attachment 321919

View attachment 321920

I see 4-6 new buttons, all of which don't apply to futures trading? Except maybe account operations, don't know what that is.

 

[rb7]

Your post does trigger a thought that I have when talking about risk in general. The way I see risk is not based on individual trades (like how much risk i need to put on a trade). I see risk as a whole from the perspective of risk of ruin (probability of blowing up).

Every strategy will have a probability of ruin. I don't care if it's a high reward to risk strategy or like mine with reverse R:R. If you are placing a trade, you're risking a portion of your capital. Whether that is 1% of your capital or 1 contract. If you lose that trade, well, you're not going to stop trading are you? You're going to continue.

Then if the next trade is a loser, you will still continue until you decide that your risk for the day/week/month is met and you stop trading OR continue trading. I'm in the camp of as long as the trading day is open, I will continue until I meet the target (minus comms)

But in either case, whether you stop early or continue trading, there is a "chance" that you can blow up your account. Example a 2 % strategy will lose 80% of the capital after about 79 consecutive losses. That is pretty low occurence. You need to kow your win/loss % and the sequence count (since we're trading futures, I take 80% loss as risk of ruin since that's the max cap futures broker will take before they auto liquidate you or stop your trading for the day).

View attachment 321897



You can lower that probability by risking small or having a max loss day like I mentioned above, but that risk of ruin is NEVER going to be zero, unless you stop trading altogether with your capital still intact.

So the way I lower my probability of ruin is trhough quick targets (amongst other things like picking the ideal time to trade, the right opening range, the right market, etc.)

I'm assuming that the combination of all those things above lead to a high probability that the strategy will survive "longer". I'm very aware that there is a chance of a blow up.

When I tested this with limited data (3 years) i didnt see the event. And now about 6 months in live, it's still functioning. But like I said in previous post, there is a price action scenario where it would kill this strategy. We'll just have to wait and see when that probabilty will hit.

I'd switch anytime from a 10-12%/month to a 5%/month strategy with an equity curve like yours, without any losing days.
But I suspect that your strategy will face a big loss at one time, whipping out all your gains (and even more). I could be wrong since I don't know any details about your strategy. But usually, strategy with small PT are prone to high win rate and high risk of ruin in the long term.
But since you backtested yours over a 3 years period, it might be different. I wish for you that it is different

It's easy to win small over a long period of time. But keeping all your gains is another ballgame (usually!).
The hard part of trading is to stay profitable over a long period of time (>2 years), without losing all past gains.

The worst part in trading once you think you are profitable, is that you never know when it will end, if it ends. It's a never ending battle. And it's a battle with yourself, not with the rest of the trading world (contrary to what most of people think).

 

[hilmy83]

I see 4-6 new buttons, all of which don't apply to futures trading? Except maybe account operations, don't know what that is.

Copy trading is very useful.

Especially when I launch my hedge fund and you're my first client per the agreement. 50k min

 

[hilmy83]

View attachment 321849

View attachment 321850

 

[Mo06]

Is it just me or is the green line below its high ?

 

[hilmy83]

Stupid algo fucked up today. I don't even care about the profit cause I just got lucky.

I feel like giving up at this point

https://www.elitetrader.com/et/threads/i-give-up.375995/#post-5856852

 

[hilmy83]

Figured out the sequence of events after reviewing the MT5 trades list and QT's event log.

Basically the algo did what its supposed to do in this order.

1. excecuted the entry in partial lots.
2. target order initiated for 1st partial lot.
3. target order filled for 1st partial lot, leaving 2nd partial lot still open
4. target order initiated again for 2nd partial lot, but it placed the order for full lots as if the whole position is still there. This cause a net of reverse position (Error)
5. Algo thinks session is over and sent out alert of ending pnl.
6. Algo placed a reverse order based on this qty of lot. The algo apparently just focus on abs qty.
7. Market executed the reverse order. So now I have reverse trade+position from step 4.
8. But I still have a naked position without stop reverse/exit
9. Found out later and exited at market for profit.

So after looking at the code and asking chatgpt, I basically decided to do a crude fix. I can't figure out how to make the algo place the correct target limit order partial position is already closed. When I looked at previous sessions where my entry is broken up, the limit targer order is always placed in full, but it's always placed after all those entries have been executed, thus no errors. Example:

So today, the target just got interrupted cause price moved quickly and executed the initial partial lot, and it kind of went downhill from there.

So the solution I have is:

// is target
            if (obj.OrderId == target_order)
            {
                // position is flat
                if (position_quantity == 0)
                {
                    if (reverse != null)
                        Log("Take Profit filled. Canceling 'reversal/stop loss' order.", LoggingLevel.Trading);
                    reverse?.Cancel();

                    //added next two lines for open positions left open
                    buy?.Cancel();
                    sell?.Cancel();

                    OnFlatten();
                    return;
                }

I added the buy/sell cancel call out. Essentially if the algo repeated step 4 again, it would just close it immediately. This can go for or against me momentarily. I figure since this is a pretty rare occurrence (hopefully), i can deal with it and the wins/loss will equalize over the long run. I just need to have this fail safe step so that I don't have a naked position once the algo says "session is done".

Back to the damn grind...God I hope this works.