To use chat, please login.
Back to contacts

length of position holding vs. pips profit vote results

Do you like this suggestion?

Yes No

length of position holding vs. pips profit Discussion

Chris
Krysztau
Aug 22 2010 at 19:31
68 posts
I think this may be key analysis to rate trading performance.
If you can show sort of chart with number of pips on Y [vertical] axis and time of holding on X [horisontal] axis.

Use of such statistics would be to see if given trader holds his 'scalps' for few hours or days to get 5 pips profit or if one cuts his losses quickly.

It would be good to see scalps held only for 5-20 minutes or only 50+ pips positions held for few hours or days [with occasional stops at BE here or there].


Please rate :)
Myfxbook (Staff)
Aug 24 2010 at 11:59
1405 posts
Sounds like an interesting statistic - thanks for the suggestion!
Steve B (stevetrade)
Aug 24 2010 at 12:02
1415 posts
Excellent suggestion for transparency
11:15, restate my assumptions: 1. Mathematics is the language of nature. 2. Everything around us can be represented and understood through numbers. 3. If you graph these numbers, patterns emerge. Therefore: There are patterns everywhere in nature.
JuanC (cacus)
Sep 01 2010 at 22:53
119 posts
+1 here
Mental note: do not abuse.
Stephanus Rensburg (stephanusR)
Sep 08 2010 at 16:35
216 posts
Another factor is the pip win to draw-down ratio. If you made 20 pips , your drawdown should be no more than 20 pips. Express this as a ratio and plot on a graph to quantify risk reward ratio. Having a 200 pip draw down isn't same as 20 . But it should take percentage of equity into account with some scaling factor. having a 200 pip draw down with one micro lot isn't the same as 1 standard lot.
Soeren (GridsForPips)
Sep 08 2010 at 18:01
205 posts
Hmm, it makes my system look worse than it is. I'm trading the all-history grid and in this case, it's expexted to have orders open for a very long time and huge DD even though it (in theory) comes back in the end for a 'small' profit. The trading gridlines and orderlotsizes is based on the full range and Equity, and there's no problem having prolonged drawdown. The danger is, when the currencies move beyond their historical boundaries, like the EURCHF, CHFJPY and the USDJPY, but my strategy involves diversification to make profits anyway, along with the widening of gridlineseparation with lower Equity, and using filter to only trade when the trend moves towards the middle for reducing DD.
I know some of you in here doesn't think there's any real profitability or 'trading' in this kind of strategy, but so far it works for me.
Until now, it seems Equity% is floating somewhere between 50 and 80, and that's fine by me, as long as the strategy pulls the 6% a month after a 'order build up' period.

If the graph is showing the gain in % versus the Risk in % and an average holdingtime for the symbol combined with magicnumber I think its alright, otherwise, I'll have to say no.

So far I'm neutral...
Always get cashback -
Chris
Krysztau
Sep 09 2010 at 10:47
68 posts
I think in both cases 200 pips in microlots vs. 20 pips in regular lots as well as in case of 'wide range strategy' you still have room to fill 'system description' space. I think if someone is examining system will certainly read creator's description.
I think this stat will filter out scammers that says they are scalpers, but in fact are holding trades for hours just for few pips profit.

Bare in mind that it's just another idea for sorting data [now I do this sort of research myself anyway].
Stephanus Rensburg (stephanusR)
Sep 09 2010 at 12:13
216 posts
Krysztau posted:
    I think in both cases 200 pips in microlots vs. 20 pips in regular lots as well as in case of 'wide range strategy' you still have room to fill 'system description' space. I think if someone is examining system will certainly read creator's description.
I think this stat will filter out scammers that says they are scalpers, but in fact are holding trades for hours just for few pips profit.

Bare in mind that it's just another idea for sorting data [now I do this sort of research myself anyway].

We must create some sort of python script that downloads the .csv file from myfxbook.com and automatically generates graphs and all sorts of data. Opensource it on google code , we will rapidly be able to zero down on the three two or even one account that is generating profits with acceptable risk .
Chris
Krysztau
Sep 09 2010 at 12:31
68 posts

stephanusR posted:
We must create some sort of python script that downloads the .csv file from myfxbook.com and automatically generates graphs and all sorts of data. Opensource it on google code , we will rapidly be able to zero down on the three two or even one account that is generating profits with acceptable risk .

I think this powerful tool is called excel spreadsheet ;)

Combined with copy and paste buttons can make miracles.
Stephanus Rensburg (stephanusR)
Sep 18 2010 at 12:07
216 posts
Another idea is to merge the length of time held factor with the pips gain : draw-down factor , if a position remains in draw down for to long and subsequently rallies, any such rally would be more due to random luck then market insight. See
https://www.myfxbook.com/community/suggestion-box/add-column-for-draw-down/52604,1

1) The length a losing position was held must be factored. If a losing position takes to long to rally then the 'random luck' factor increases. Each trade is assigned a 'random luck' factor weighted for equity used, meaning a standard lot will have more impact on this factor than a micro lot. The overall account is assigned this averaged out 'random luck' factor.
2) Length a winning position was held expressed as single number weighted for all trades. Holding a micro lot for one year on an equity of $10000 should have a very small impact.


The time it took for a trade to rally and the draw down incurred must be combined and rolled into a single number to quantify the account and differentiate it from gambling as opposed to market insight.

Trading periods should also be tagged as either being in a trend, counter trend or break-out. This is done to differentiate accounts that traded in the direction of the trend only , but not yet in a counter trend. A trader must be able to trade a ranging and trending market.
Chris
Krysztau
Sep 18 2010 at 16:11
68 posts
Stephanus are you sure your proposals fit this topic?
I am sure your ideas are great for some data analisys, but I personally don't ind them fitting into two dimension chart, where one dimension is a time and other is a profit.

It's obvious that not a single trader will obtain correlation close to 1.

In my opinion the main use of such statistic is separation scalpers from swings, long terms and 'scamer-scalpers'.
Stephanus Rensburg (stephanusR)
Sep 18 2010 at 18:52
216 posts
Krysztau posted:

It's obvious that not a single trader will obtain correlation close to 1.

In my opinion the main use of such statistic is separation scalpers from swings, long terms and 'scamer-scalpers'.

It is up to the evaluator of the system. Myfxbook.com won't implement it though because the site is driven by ads. One would have to have an automation script that automatically downloads all the top 100 CSV files and then run an excel, python script to generate this number. Same problem with collective2.com, they don't present the trading stats in such a way that the true risk: reward ratio is clear. Risk:reward quantifies the long term sustainability of a trading method, its statistical edge.

If we can reduce a jumbled mass of numbers to a single ratio we would make significant strides in enabling investors to make an informed investment decision as opposed to a gambling decision.

Scalping, swing trading methods would now be quantified in the same manner for their statistical edge. In other words if you scalp for 1pip, then your draw-down had better be 1pip or what ever draw down one would find acceptable. Lets presume your win rate is 90%, pip gain is only 5, with a draw down of 10 and SL of 20. If the win rate is 60% but the pip gain is 50 with a draw down of 10 and SL of 20 then a better risk profile is demonstrated. We need to combine stated SL at entry of trade, average TP, draw down and period held to compensate for the 'random luck' element in trading.

A system must state the SL at the entry of the trade or we wind up with a situation of multiple open positions that the trader hopes will eventually close for a profit , incurring large draw downs in the process. Any subsequent gain would then be ascribed to the 'random luck' factor.

Most would agree that it is remarkable how complicated it can be to find a system that generates an engineered return on equity and not a gambled return on equity.
Please login to comment .