MFI CROSS strategy to test and improve on Wall Street 6 minutes
-
HI,
Here an idea to use MFI that I consider very interesting.
The strategy (ONLY LONG) waits for a cross of a MFI fast and slow.
Moreover the average must always go up in the last period.
The system perform well on wall street (6 minutes) and has a little DD.
It would be nice if anyone can tests it on a frame of time greater than one year (as I can only do).
Let me know….
bye
//----------------------------------------------------------------------------------------------- //WALL STREET // 6 MIN. DEFPARAM cumulateorders = FALSE CONDAVG = 1 AVG = Average[S](close) FOR C = 1 TO F IF AVG[C-1] > AVG[C] THEN IF CONDAVG = 1 THEN CONDAVG = 1 ENDIF ELSE CONDAVG = 0 ENDIF NEXT TIMEWORK = 000000 TIMESTOP = 240000 F = 20 S = 200 MFIFAST = MoneyFlowIndex[F] MFISLOW = MoneyFlowIndex[S] BASEMFI = MFISLOW[0] > MFISLOW[F] AND MFISLOW[0] > MFISLOW[ROUND(F/2)] AND MFISLOW[0] > MFISLOW[ROUND(F/8)] CONDMFIL = MFIFAST[0] CROSSES OVER MFISLOW[0] //TRAILINGSTART = 50 * PIPSIZE TRAILINGSTART = (close/((S*2)+(F*2))) * PIPSIZE AMNT = TRAILINGSTART * POINTVALUE * COUNTOFPOSITION //GRAPH AMNT SIZEBUY = 1 IF CURRENTTIME >= TIMEWORK AND TIME <= TIMESTOP THEN if not onmarket THEN IF CONDMFIL AND BASEMFI AND CONDAVG THEN BUY sizebuy SHARES AT highest[F](HIGH) STOP endif ENDIF endif IF ONMARKET THEN PERFP = CLOSE - TRADEPRICE PERFN = (PERFP / TRAILINGSTART) TRAIL = (TRAILINGSTART/PERFN) *POINTVALUE * COUNTOFPOSITION IF PERFP > TRAILINGSTART THEN SET STOP $TRAILING TRAIL ELSE SET STOP $TRAILING AMNT ENDIF IF OPENTIME > 213000 AND DayOfWeek = 5 THEN SELL AT MARKET ENDIF ENDIFHI GraHal,
I modified the code as suggested and results (on 100K bars) are good but with less protection during the bear market phases.
I have to test it on demo in order to check if actually I can see the stop loss line moving (but I quite sure that is not so – sorry if I insist on it – maybe I am wrong…)
The reason why Nicolas couldn’t see on 200k is unknown to me, I can’t understand it as I can see only 100k (if I have correctly understood your affermation).
This the code modified:
//----------------------------------------------------------------------------------------------- //----------------------------------------------------------------------------------------------- //WALL STREET // 6 MIN. DEFPARAM cumulateorders = FALSE CONDAVG = 1 AVG = Average[S](close) FOR C = 1 TO F IF AVG[C-1] > AVG[C] THEN IF CONDAVG = 1 THEN CONDAVG = 1 ENDIF ELSE CONDAVG = 0 ENDIF NEXT TIMEWORK = 000000 TIMESTOP = 240000 F = 20 S = 200 MFIFAST = MoneyFlowIndex[F] MFISLOW = MoneyFlowIndex[S] BASEMFI = MFISLOW[0] > MFISLOW[F] AND MFISLOW[0] > MFISLOW[ROUND(F/2)] AND MFISLOW[0] > MFISLOW[ROUND(F/8)] CONDMFIL = MFIFAST[0] CROSSES OVER MFISLOW[0] //TRAILINGSTART = 50 * PIPSIZE TRAILINGSTART = ROUND((close/((S*2)+(F*2)))) * PIPSIZE TRAILINGSTEP = ROUND(TRAILINGSTART / 4) //AMNT = TRAILINGSTART * POINTVALUE * COUNTOFPOSITION //GRAPH AMNT SIZEBUY = 1 IF CURRENTTIME >= TIMEWORK AND TIME <= TIMESTOP THEN if not onmarket THEN IF CONDMFIL AND BASEMFI AND CONDAVG THEN BUY sizebuy SHARES AT highest[F](HIGH) STOP endif ENDIF endif //reset the stoploss value IF NOT ONMARKET THEN newSL=0 ENDIF //manage long positions IF LONGONMARKET THEN //first move (breakeven) IF newSL=0 AND close-tradeprice(1)>=trailingstart*pipsize THEN newSL = tradeprice(1)+trailingstep*pipsize ENDIF //next moves IF newSL>0 AND close-newSL>=trailingstep*pipsize THEN newSL = newSL+trailingstep*pipsize ENDIF ENDIF IF newSL>0 THEN SELL AT newSL STOP ENDIFModerator’s edit: code edited to apply the correct PRT code format, please use “insert PRT code” button for future occurences (last one at end of first line of tools in message editor), thanks
DO you test in Demo account ???
seems interest code , but I think could be more secure with the stop loss set up in the code ….
FOllow the backtest with 0.2 % and with 1.7 % Stop loss
Actually the lack of complete 200k bar history is a known bug in PRT. I’ve raised it a couple of years back with them but it hasn’t been addressed. And also, there is volume history on the Dow going back further than 12mths, in fact it starts on Jul 2016. The reason there is no more data on 200k bars than 100k bars above is because of a weird bug in PRT that only calculates the full 200k bars based on “standard” time units ie those that you get on the drop down menu. At least that’s my theory anyway. See below.
As you see from the first image there is no more data on the 100k chart compared to 200k chart for the 6m TF
On the second image there is twice as much data in 200k units for the 5m period as there is for 100k units – as expected (its the same with 10m tf as well)
Also, I don’t think there is any correlation between the units shown in the drop down menu and the actual time scale on the charts.
Eg. for the 6m tf – 6m x 200,000 = 1.2m / 60/ 24 = 833 days /365 = 2.28 years ie 2yrs and 3mths approx. And as you can see from the history we get only around half that. In fact in reality it should be around 15% more history than even that figure due to the fact that Saturday has no units. I think you just have to accept this limitation in PRT though it’s a shame that if the data is there and is not being made available.
GraHal thanked this postHI JR1976,
I’m cheking several version of the strategy(on demo) in order to find the best one.
Anyway I keep in consideration your suggestion and I’m preparing the proper code to manage the SL with the code provided by Nicolas.
Rusults with the version I posted (in attached) are quite good even on DD as your result in % on SL.
I’ll let you know further develops..
thx for you time.
bye
- You must be logged in to reply to this topic.
Summary
This topic contains 38 replies,
has 8 voices, and was last updated by 
7 years, 3 months ago.
Topic Details
| Forum: | ProOrder: Automated Strategies & Backtesting |
| Language: | English |
| Started: | 03/12/2018 |
| Status: | Active |
| Attachments: | 16 files |
