TP Tree Method | Page 3 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Is There Any Simplest Method To Success In Forex? 63 replies

bo7a method... method for GBP/JPY 205 replies

(Method) Trading Systems 1 reply

Inquiry of 400:1 leverage method 11 replies

Simple trading method 49 replies

Attachments: TP Tree Method
Exit Attachments

TP Tree Method

  • #41
  • Jul 17, 2008 8:41am Jul 17, 2008 8:41am
Quoting Ronald Raygun
If that's the case, then it's not your risk which gets impacted, but instead it's your win rate. If--after every loss--you increase the TP to compensate, you also reduce your probability of making a profit.



Well, by your system, every time you get stopped out, the TP increases. This will be especially painful if your system is caught in a ranging period. You'll have multiple stop-outs and an almost impossibly high TP which ultimately dwarfs my TP requirements.

Personally, I'm thinking of dreamliner's comment and I think he/she is right. What edge do we have? I suppose I'll just combine my pyramiding with something with a breakout-style system. That seems to be the best option for me.
Yes, I know. I've already pointed out the crux of the system several times. I said there would need to be tweaking.

That's why I want to try martingaling the first few trades. That would allow the TP to not increase for the first few trades. The success of tkimbles methods have me intrigued to believe that by martingaling and having a set TP level, we can increase the probability of success. Besides, the system does have a chance at profiting in a ranging market depending on how large the range is. It will profit in a trending market. I don't think you can say a lot of other systems have that kind of potential.

Eventually I want to try an 'uncle' point where it makes sense to just reset the system while still having a positive expectancy. Perhaps the ranges can't be too small as the system may need to only play the long term market movement and not in the market noise. Perhaps the best bet is to not martingale and just trade a straight system that institutes the uncle point and trades longer time frames. Trading it on multiple ranges with multiple currencies should create the flurry of activity to keep the profit rolling in.

I never said, 'this system works as described.' I think the idea merits trying to figure out a method that does work.
  • #42
  • Jul 17, 2008 9:17am Jul 17, 2008 9:17am
Quoting permanentjaun
I would hope someone is kind enough to put this in to code. Without any indicators it might be easy enough, although the money mangement may be tricky.
Matt, thanks for sharing!

I wrote an EA which trades like FXTradePro's semimartingaling sequence. Should be easy to modify it. Will do it.

Regards - Xaron
  • #43
  • Jul 17, 2008 10:15am Jul 17, 2008 10:15am
Quoting Xaron
Matt, thanks for sharing!

I wrote an EA which trades like FXTradePro's semimartingaling sequence. Should be easy to modify it. Will do it.

Regards - Xaron

Thank you. I look forward to seeing it.

In the meantime I have another idea. As has been pointed out several times, the problem of the system is if the TP level gets out of hand. One other idea to throw into the mix is to take cuts along the way. So say we're trading a 50/50 system. Have that hard 50 SL, but institute a trailing stop of 100 pips.

Presume we've been stopped out several times in our martingaling system and we're looking for 200 pips. With the 100 trailing stop loss, price retraces from +200 and stops that one specific trade out at +100. We can either decide to reset the system and be happy the total loss has been reduced by half. Or we can say, ok on this series we were looking for a 200 pip profit, but have reduced that to 100. On the next trade keep the position size equal to the last trade, but reduce TP to 100. The hard SL would be 50, and another 100 pip trailing SL.

Then, that trade series does not reset to the original TP/SL and position size until we've banked the amount of total pips we were looking for. That could help things if the TP level starts to get large.


I'm just throwing ideas out there for discussion.

Matt
  • #44
  • Jul 17, 2008 1:10pm Jul 17, 2008 1:10pm
Hi all,

here comes the first version of the EA.

My first backtests weren't so great, but I've just used these settings:

Pair: EUR/USD
SL/TP: 20/20 for the first 5 trades (size: 0.01, 0.02, 0.04, 0.08, 0.16)
After that always 0.16 lot but doubling TP...

I stopped after TP reached 160 pips in the 8th level, more doesn't make any sense I think.

You can add up to 20 levels if you want and set the size as well as the TP and SL for each of these levels.

You can use a price entry or a time entry or a manual one...

Regards - Xaron
Attached File(s)
File Type: zip AdvancedSAR_v1_0.zip   8 KB | 445 downloads
File Type: ex4 AdvancedSAR_v1_0.ex4   30 KB | 429 downloads
  • #45
  • Jul 17, 2008 2:46pm Jul 17, 2008 2:46pm
Quoting Xaron
Hi all,

here comes the first version of the EA.

My first backtests weren't so great, but I've just used these settings:

Pair: EUR/USD
SL/TP: 20/20 for the first 5 trades (size: 0.01, 0.02, 0.04, 0.08, 0.16)
After that always 0.16 lot but doubling TP...

I stopped after TP reached 160 pips in the 8th level, more doesn't make any sense I think.

You can add up to 20 levels if you want and set the size as well as the TP and SL for each of these levels.

You can use a price entry or a time entry or a manual one...

Regards - Xaron
I'm at work and not able to view/test this yet. One thing that might be skewing your results however is how you mentioned this:

"SL/TP: 20/20 for the first 5 trades (size: 0.01, 0.02, 0.04, 0.08, 0.16)
After that always 0.16 lot but doubling TP..."

The system does not require TP to be doubled. It only increases by the size of your SL. Here is how the first 10 trades should play out if you are using a 20/20 TP/SL.

Trade TP--- SL--- Lot Size

1 --- 20--- 20--- 1
2 ---20 ---20--- 2
3 ---20 ---20--- 4
4 ---20 ---20--- 8
5 ---20 ---20 ---16
6 ---40 ---20 ---16
7 ---60 ---20 ---16
8 ---80 ---20 ---16
9 ---100---20 ---16
10---120---20 ---16

As you can see this should drastically change your results as even at the 10th level you're still two more trades away from requiring a 160 pip move. In your current model then the 9th trade would require a 320 pip move. Under the model I just presented, a 320 pip movement would not be required until our system has been stopped out 21 times in a row. That should create a large discrepancy in the results if that was not coded into the EA correctly.

I also have one more request if you don't mind. Can you take out the function for now that says, "after so many trades in a series, stop that series and start over,"? I want to try testing the system on a longer time frame such that it might take a week or two to complete the series. Weekly candles can easily reach 500 pips in length, and therefore I would not mind giving the system a chance, especially if it is only looking for a 160 pip move.

That is more of a preference thing for me. You might only want your trades to occur short term. I'm fine with just letting it run. I'd like to at least test it at this point.

Thanks for putting this together. I may not get to it tonight, but I can't wait to start playing.

Matt
  • #46
  • Jul 17, 2008 2:57pm Jul 17, 2008 2:57pm
Ah.... my fault. Sorry.
  • #47
  • Jul 18, 2008 7:26pm Jul 18, 2008 7:26pm
I haven't taken a look at the file yet, will do so shortly. I have no coding experience so I don't know what it would take to make the changes that need to be made to trade the system as described.

I think it would be worthwhile to eventually try the 10/100 system. Really, the smallest range would be interesting. I'll have to think about the most efficient way to grab the pips. I was afraid of using this type of scalping range because you can't martingale, but I was an idiot. We can martingale with the 10/100 system. The martingaling would just take longer to initiate.

Here is a breakdown of lot sizes per trade.

Trade --- TP/SL --- Lot Size
1---------10/100------1
2--------110/100------1
3--------110/100------2
4--------110/100------4
5--------110/100------8
6--------110/100------16
7--------210/100------16
8--------310/100------16
9--------410/100------16
10-------510/100------16

That should work. What would be an interesting study would be to determine the 'uncle' point that I've mentioned before.

Uncle point = the point at which the TP level is so large that we can't bear to wait for it to complete that series. We hit SL and reset the series.

What I think is interesting is that the scalping trades might hit many TP's in a row. Does that necessarily mean we're able to compound profits quicker, finely collect the pips to get more, and thus make it easier for us to lower our uncle point? I'll give an example to show what I mean.

For one, the 10/100 type of system trades very similarly to a 100/100 system once that first SL is hit. But say we are in a trade that goes 85 pips in the direction of our trade. Disregarding spread for this example, the 10/100 system could effectively collect 80 pips. A 30/30 system would only be able to collect 60 pips since it would have required price to have moved 90 pips to bank that last trade. So that is my point, is trading the scalping version always better?

I think the answer is determing what affect paying the spread many times over has on the profit. I get the feeling it will have a huge effect. If spread is 2 pips in the above example, we would have paid 16 pips in spread to reduce the total pips on the trade to 64. I'm not sure if that's the correct math involved.

Any ideas people?
  • #48
  • Jul 21, 2008 1:30am Jul 21, 2008 1:30am
Quoting permanentjaun
I haven't taken a look at the file yet, will do so shortly. I have no coding experience so I don't know what it would take to make the changes that need to be made to trade the system as described.
You don't need to change anything. Everything can be changed from the outside, there are a lot of external variables.

Regards - Xaron
  • #49
  • Jul 21, 2008 4:34pm Jul 21, 2008 4:34pm
Great. I'll hopefully get to it tonight. Have you played around with it on the corrected settings then? Any results good or bad?
  • #50
  • Jul 21, 2008 4:41pm Jul 21, 2008 4:41pm
hate to sound cynical, but i read this thread over the weekend, and see the system blowing up eventually.

you can get caught in an extreme ranging market.

i recommend waiting for a few "bounces" before entering... in other words, start the sequence later than just a random entry.

-t
  • #51
  • Edited 5:35pm Jul 21, 2008 4:57pm | Edited 5:35pm
Yes I see that too. I like the word, 'eventually' though. That ties along with the 'uncle' point. How many bounces can we soldier through until we say we'd be better off just resetting the system? You'd have to define an extreme ranging market too. What would an extreme ranging market look like that would be able to stop out a 30/30, 60/60 and 90/90 system all at the same time to the point that you don't want to continue the trade series on all three systems?

I'm not necessarily looking to build a system that works all of the time, but only most of the time. And by martingaling, expanding tp, etc etc, I'm looking to reduce the number of scenarios the system can break down from.

Your last thought makes me think you'd rather see this transformed back into tkimbles version of waiting for some signal; in their case a breakout. Personally I think that is just slowing down the time it takes to experience one of those blow-up catastrophes, but also profiting less frequently. I'm willing to try both.
  • #52
  • Jul 21, 2008 5:36pm Jul 21, 2008 5:36pm
dude, it's not so much about waiting for a signal as it is about avoiding drawdowns.... waiting for a few bounces across the range (that you WOULD have taken) gives you a better chance of catching your breakout before the uncle point, because you are letting a few losses pass you first.

these are just raw thoughts and not recommendations.

Quoting permanentjaun
Your last thought makes me think you'd rather see this transformed back into tkimbles version of waiting for some signal; in their case a breakout. Personally I think that is just slowing down the time it takes to experience one of those blow-up catastrophes, but also profiting less frequently. I'm willing to try both.
  • #53
  • Jul 21, 2008 6:14pm Jul 21, 2008 6:14pm
You'll have to explain how we'd do that. How would you know how many bounces there will be or that one that you pass up is actually the breakout formation?
  • #54
  • Jul 21, 2008 6:16pm Jul 21, 2008 6:16pm
clearly you'd miss some easy winners, but is it worth missing a few small take profits vs. blowing the account (or hurting it severely.... IE wiping out all gains, etc?)

there is no magic bullet. you could run some backtests and find out how many skipped bounces would have historically kept the account safe.

again, just thinking out loud....
  • #55
  • Jul 21, 2008 6:33pm Jul 21, 2008 6:33pm
I guess the important question for your model would be the rules on reentry. What happens when you skip a bounce and price moves outside the original range, ie we skipped a trade that would have hit TP?

Also, I'm not sure how to answer the question of , "is it worth missing a few small TP's?" Technically you are correct in that TP is larger as the trades progress, BUT we are only banking a certain amount of pips on each trade. If we're looking for 10 pips profit on the first trade, then we're also looking for 10 pips profit on the last trade. The percentage of our TP that goes towards net profit decreases with each loss. In that aspect I see every TP opportunity as critical. We can play with it.

I'd also have to wonder how possible it is to blow an account with the money management I've outlined. I think in a previous post I showed that even with ten losses we've only lost 11.1% of our original account balance if we martingale the first 5 trades. This is still considerably lower than traditional MM practices. The only thing that would increase drawdown in the system to a traditional MM system is if we consistently go to 10 SL's or more. Drawdown can then reduced if we run this across several time periods and currencies that are not correlated and therefore produce more frequent winners while that other system is going to 10 losses. Any thoughts on that?
  • #56
  • Jul 21, 2008 6:39pm Jul 21, 2008 6:39pm
i don't think you understand what i mean when i say "skipped bounces"

what i'm saying is, let the market settle into a resistance level before entering the very first trade. in other words, had you placed your first hedge at the very start of the range (resistance level) you would have incurred several losses.... but by skipping the first few bounces before entering the first trade, you are counting on the market breaking away from resistance soon.

BTW, 10 losses in a row w/ a 10 pip stop is NOTHING for even EURUSD. the pair can bounces between 10 pips (in a 30 pip range) 100 times, even if it doesn't appear so on the chart. keep in mind, you can't see the bounces inside of the 5 minute candlestick... sometimes even the 1 minute.
  • #57
  • Jul 21, 2008 6:51pm Jul 21, 2008 6:51pm
Oh I thought you meant skipping trades after you had begun a trade series. I see what you're saying then. What would be the difference between doing that and waiting for a signal of a breakout from that range then? I was trying to build the system to be blind and be always in, since I wanted to avoid the common problem with other systems of requiring to predict the market to some extent. Maybe giving this system a little bit of an edge would be worth it though.

BTW, yes I know currencies can have those 10 pip bounces many times. That is why I'd like to play around with 30/60/90, maybe 120 pip ranges. It would slow the system down, but we'd be avoiding the stupid market noise. I can wait if it means improving the efficiency of the system. Just like how I mentioned earlier too, playing the system on multiple time frames over 2-3 currencies should produce trades consistently even on a larger time frame.

Your MA optimization thread certainly comes to mind, but it would be hard to implement a study like that with this method.
  • #58
  • Jul 24, 2008 8:31pm Jul 24, 2008 8:31pm
Xaron,

You've got a PM.

Also, I'm forward testing this on a demo account now. For some reason the EA seemed to skip the first trade and go straight to sequence two with the lot size at .02 even though price may have only swung 3-4 pips while I'm testing at a 20/20 TP/SL setting. Any idea why that would happen?
  • #59
  • Jul 31, 2008 11:40am Jul 31, 2008 11:40am
Quote:
Originally Posted by Leonlorenzo http://www.forexfactory.com/images/buttons/viewpost.gif
Great/interesting idea youve got here.

Certainly new ways to look at things.

But previously you say 'martangle increases risk', remember... you are also increasing risk by having a larger TP. The larger the TP the more chance of a retrace that will nock you out.

Just a thought.


Leon


Quoting permanentjaun
Certainly, I agree. It does reduce your probability of hitting TP, which is, as mentioned, the crux of the system. We have to find ways to reduce the probability of that happening. In the same light, someone could trade the other variation of the system I mentioned where TP is 10 and SL is 100. You'll likely hit many TP's in a row and then eventually hit SL.
Quoting permanentjaun

I'm not so sure how far we can say we're increasing the risk of the system by icnreasing TP though. Risk is a touchy word that has many definitions. I define it as how much we have to lose on the trade. So lot size and size of SL is what I look at in relation to account size.



I think it is very difficult to calculate "probability" in trading (not like the casino gambling games) because it involves many more factors which are keep on changing in relative to each other in trading.
For instance, the concern of widening the TP will increase the risk as there will be more chances to be stoped out by a retrace is generally true when without consider other factors or provided all other factors remain unchanged. But, if other conditions change, increase the TP will be either reduce or increase the risk.
As such, I think that the trading system/method explored in this thread, when we increase the TP after a number of SAR trades started from a fixed price levels (Long or Short) may not increase the "trading" risk. This is because there is an increased chance for the price action to move farer above or under from the fixed entry level to hit the TP (Volatility maybe going to increase because some "times" has passed since the 1st entry.

I am not sure whether there is a math formula/method to calculate the probability of something in trading, just because the factors needed in the calculation just keep on changing and are changeable...


Just a thought!
  • #60
  • Aug 1, 2008 11:56am Aug 1, 2008 11:56am
Bingo! You got it. Price will more often than not move away from your entry price and to larger and larger distances the more time you give the trade.
0 traders viewing now