My algo trading daily performance

cloudTrader

Well-Known Member
Ya, I was also shocked to see that kind of slippage ! What could be the reasons for such huge slippage value ?

Could it be possible that there is some lag in the datafeed etc., such that till the time the trade trigger gets generated in mechtrader VPS, the real market might have already move away from that price, thereby causing the slippage.

I experienced such issue with my esignal datafeed few years back, when there was considerable lag in the price being showed in the tick charts based on esignal, and the actual price in the broker terminal where I was punching in the orders. It noticed it to be even 3-4 seconds late sometime, although esignal is a professional and authorized data feed.

I have no idea, if that could be the case with mechtrader as well, but I am just guessing.

Hopefully, he has done proper research in this direction and is aware of the real reasons behind such huge slippage.

Thanks and regards
TradeOptions bro ,

Are you using E-signal data feed as of now? If not E-signal then which data feed are you using as of now?

PS: Your PM Inbox is full to receive any new messages. I wanted to send a message but could not.
 

TradeOptions

Well-Known Member
TradeOptions bro ,

Are you using E-signal data feed as of now? If not E-signal then which data feed are you using as of now?

PS: Your PM Inbox is full to receive any new messages. I wanted to send a message but could not.
yes brother, I am still using esignal only. You may directly contact me at 5555vegas @ gmail.com

Regards
 

jallanankit

Well-Known Member
Hi all,




Everything apart from the performance of the model was under control and within limits.


MT
Hello Mechtrader..

The intention to double the position size was well planned and I acknowledge your decision. But with my experience I will disagree on your point of cutting the position to half and one-fourth with system drawdown increases.

My version on the same is as follows:
1. When the system DD increases, so does the possibility of reversal. Bcoz that is how the sys DD is.. In the backtest pls refer to such periods..

Pls consider the following..

Let us consider your backtest from 2011 to 2015 end. period of 5 years.
Say the Max Sys DD in 5 years was 100000/- (100%)

Again in June 2012, the system got hit and max DD in 2012 was 80000/- 80% of the max DD in the backtested period of 5 years!! PLs see the reports of July, and Aug.. See how doubling the qty may have worked wonders for the user..

I mean to say that when the system is at its worst(meaning when the sys is near to the max dd level, that is the time to double/increase the position size.. My idea is to double the qty whenever the system DD reached 80-90% level of the sys DD...

Pls comment on my reasoning and let me know where do u disagree and put forth your reasons for supporting why position size should be reduced when DD increases..

All the very best for next two series. April and May..
Just believing in your system and wishing to see abnormal profits flow on account of my understanding of the backtest report logic..

(NB: even the system which I am trading managed only 2.5% in Feb and -1% in Mar and my Sys DD is close to 90%) I m considering this as one of the best times to increase (double) my qty.

Regards,
Ankit
 

mechtrader

Well-Known Member
I will try to explain my reasoning for my cutting my position size in half.

1) I always believe that the worst drawdown of a strategy is always in the future. If we test a strategy from 2010 - 2012 the worst DD is eg. 3l and if we test the same strategy from 2010-2015 the worst DD is 5l. The strategy is very good and has passed all our robustness criteria to go live. What would have happened if we went live with the strategy in 2012 end. What we can see from our backtest is that the strategy has max DD of 3l. If that got breached in 2013, what could we have done? Shall we stop trading it? Shall we double the position size? Certainly the strategy was not broke and it did well in next 2 years. Even the strategy can give >150% of the backtested DD and still be profitable. The strategies dont die suddenly, they pass through a lean period and have a slow death.

I personally give buffer to all my strategies till 150% of the backtested DD and keep faith in them.

Now all this was considering only one live strategy. Let me put forth some views about portfolio of strategies.

2) I run 5 strategies. Lets assume the DD of all 5 individually is 4l. So, the DD of all 5 combined the worst case is 20l. During my portfolio testing i found it to be close to 12l. Why was that so? because some performed better when some didnot. This is the beauty of portfolio trading. Now the problem is how can we arrive at the max DD of the portfolio. The answer is MC simulations. Although I have not done it but believe the average of MC simulations will be around 15l. Also, hypothetically the worst DD is 20l, but giving each strategy a buffer (which we learnt above) the DD reaches 30l. So our worst case scenario can be 30l hypothetically.

Now, in portfolio trading I can not discard any of the strategy (beacuse they are all under their 150% limit) and still be hitting a new worst DD. What can i do in this situation? I donot know where can it stop? Shall i go with double position size ?

No, i wont do that, If i were running a single strategy, i might have done it, but not in portfolio. I believe the best measure is to let the DD pass. I believe that with my current method of handling of DD i might miss out on some profits, but capital conserving is my first priority.

Lets see how it goes. It certainly will be a learning experience.

I have borrowed some of my ideas from this mans blog. This article is somewhat related. http://www.systemsontheroad.com/#!H...ING-SITUATIONS/c18q6/56ac81780cf231794c631bb0

Thanks,
MT
 

DSM

Well-Known Member
MechTrader,

Just a question. Assume that you are fully in position in a ceratain direction (long or short) by 11AM or 12PM. What if there is a great change in sentiment/market direction and the market zooms in the opposite side? Is such a scenario considered to exit out of trades to cut off positions/loss on basis of change in trend/direction earlier than the closing time?

Thanks.

I will try to explain my reasoning for my cutting my position size in half.

1) I always believe that the worst drawdown of a strategy is always in the future. If we test a strategy from 2010 - 2012 the worst DD is eg. 3l and if we test the same strategy from 2010-2015 the worst DD is 5l. The strategy is very good and has passed all our robustness criteria to go live. What would have happened if we went live with the strategy in 2012 end. What we can see from our backtest is that the strategy has max DD of 3l. If that got breached in 2013, what could we have done? Shall we stop trading it? Shall we double the position size? Certainly the strategy was not broke and it did well in next 2 years. Even the strategy can give >150% of the backtested DD and still be profitable. The strategies dont die suddenly, they pass through a lean period and have a slow death.

I personally give buffer to all my strategies till 150% of the backtested DD and keep faith in them.

Now all this was considering only one live strategy. Let me put forth some views about portfolio of strategies.

2) I run 5 strategies. Lets assume the DD of all 5 individually is 4l. So, the DD of all 5 combined the worst case is 20l. During my portfolio testing i found it to be close to 12l. Why was that so? because some performed better when some didnot. This is the beauty of portfolio trading. Now the problem is how can we arrive at the max DD of the portfolio. The answer is MC simulations. Although I have not done it but believe the average of MC simulations will be around 15l. Also, hypothetically the worst DD is 20l, but giving each strategy a buffer (which we learnt above) the DD reaches 30l. So our worst case scenario can be 30l hypothetically.

Now, in portfolio trading I can not discard any of the strategy (beacuse they are all under their 150% limit) and still be hitting a new worst DD. What can i do in this situation? I donot know where can it stop? Shall i go with double position size ?

No, i wont do that, If i were running a single strategy, i might have done it, but not in portfolio. I believe the best measure is to let the DD pass. I believe that with my current method of handling of DD i might miss out on some profits, but capital conserving is my first priority.

Lets see how it goes. It certainly will be a learning experience.

I have borrowed some of my ideas from this mans blog. This article is somewhat related. http://www.systemsontheroad.com/#!H...ING-SITUATIONS/c18q6/56ac81780cf231794c631bb0

Thanks,
MT
 
If you reduce your positions, won't it lead to extension of your DD and resulting into systems not stopping where they are suppose to stop? Or in case of recover the EC won't be as per the original system?


I will try to explain my reasoning for my cutting my position size in half.

1) I always believe that the worst drawdown of a strategy is always in the future. If we test a strategy from 2010 - 2012 the worst DD is eg. 3l and if we test the same strategy from 2010-2015 the worst DD is 5l. The strategy is very good and has passed all our robustness criteria to go live. What would have happened if we went live with the strategy in 2012 end. What we can see from our backtest is that the strategy has max DD of 3l. If that got breached in 2013, what could we have done? Shall we stop trading it? Shall we double the position size? Certainly the strategy was not broke and it did well in next 2 years. Even the strategy can give >150% of the backtested DD and still be profitable. The strategies dont die suddenly, they pass through a lean period and have a slow death.

I personally give buffer to all my strategies till 150% of the backtested DD and keep faith in them.

Now all this was considering only one live strategy. Let me put forth some views about portfolio of strategies.

2) I run 5 strategies. Lets assume the DD of all 5 individually is 4l. So, the DD of all 5 combined the worst case is 20l. During my portfolio testing i found it to be close to 12l. Why was that so? because some performed better when some didnot. This is the beauty of portfolio trading. Now the problem is how can we arrive at the max DD of the portfolio. The answer is MC simulations. Although I have not done it but believe the average of MC simulations will be around 15l. Also, hypothetically the worst DD is 20l, but giving each strategy a buffer (which we learnt above) the DD reaches 30l. So our worst case scenario can be 30l hypothetically.

Now, in portfolio trading I can not discard any of the strategy (beacuse they are all under their 150% limit) and still be hitting a new worst DD. What can i do in this situation? I donot know where can it stop? Shall i go with double position size ?

No, i wont do that, If i were running a single strategy, i might have done it, but not in portfolio. I believe the best measure is to let the DD pass. I believe that with my current method of handling of DD i might miss out on some profits, but capital conserving is my first priority.

Lets see how it goes. It certainly will be a learning experience.

I have borrowed some of my ideas from this mans blog. This article is somewhat related. http://www.systemsontheroad.com/#!H...ING-SITUATIONS/c18q6/56ac81780cf231794c631bb0

Thanks,
MT
 

TradeOptions

Well-Known Member
MechTrader,

Just a question. Assume that you are fully in position in a ceratain direction (long or short) by 11AM or 12PM. What if there is a great change in sentiment/market direction and the market zooms in the opposite side? Is such a scenario considered to exit out of trades to cut off positions/loss on basis of change in trend/direction earlier than the closing time?

Thanks.
This is a very common market scenario. The positions are exited based on stop losses or trailing stop losses as usual.

MT
So basically NO CHANGE at all in the Mechanical Approach, even if the market shows "a great change in sentiment/market direction and the market zooms in the opposite side". You would wait for the stops to get hit for getting out of those position.

The most probable reason, as to why you choose to do it like this, could be because you do not want to involve subjectivity into it, and moreover the market direction might change again, once you exit the positions subjectively.

Regards
 
MechTrader,

Just a question. Assume that you are fully in position in a ceratain direction (long or short) by 11AM or 12PM. What if there is a great change in sentiment/market direction and the market zooms in the opposite side? Is such a scenario considered to exit out of trades to cut off positions/loss on basis of change in trend/direction earlier than the closing time?

Thanks.
One Hedge fund in newyork lost 400 million$ in 5 minutes, just of bug, they couldn't close position.
It's very crucial to have regularly updating codes and make it more secure.