⚡️ Weighted Signal: PVT - Price Volume Trend

[andif888]

This adds a "Price Volume Trend Indicator" to the strategy according to this implemenation at https://www.tradingview.com/script/3Ah2ALck-Price-Volume-Trend/

[Rikj000]

First of all, thank you! 🙏
But it's a bit unclear to me which test results are which, some direct BackTest results here would be the easiest for us!

• So I'm assuming the older .fthypt file would be a vanilla test and the newer .fthypt file a test with the new PVT indicator implemented?
• Or is this 2 tests of 1 process (Run 1 & Run 2) with the PVT indicator implemented?

In case of the latter then I'll have to ask you for a vanilla test (without the PVT indicator, but using the same --random-state so we can compare before/after on an identical run)

[andif888]

commit: dc3142a

.fthypt file without pvt -> strategy_MoniGoManiHyperStrategy_no_pvt.fthypt

.fthypt file with pvt -> strategy_MoniGoManiHyperStrategy_with_pvt.fthypt

command line:

freqtrade hyperopt -s MoniGoManiHyperStrategy \
-c ./user_data/mgm-config.json \
-c ./user_data/mgm-config-private.json \
--hyperopt-loss WinRatioAndProfitRatioLoss \
--spaces all \
-e 300 \
--random-state 2 \
--timerange 20210521-

highlevel results:

without pvt

+--------+---------+----------+------------------+--------------+-------------------------------+-----------------+-------------+                            
|   Best |   Epoch |   Trades |    Win Draw Loss |   Avg profit |                        Profit |    Avg duration |   Objective |
|--------+---------+----------+------------------+--------------+-------------------------------+-----------------+-------------|
| * Best |   2/300 |        4 |      0    4    0 |        0.00% |                            -- | 0 days 05:44:00 |          -0 |
| * Best |   4/300 |       70 |     31   20   19 |        0.92% |        29.037 USDT    (5.81%) | 0 days 11:19:00 | -2,854.78477 |
| * Best |   9/300 |       67 |     43    2   22 |        0.80% |        24.240 USDT    (4.85%) | 0 days 08:20:00 | -3,453.67246 |                           
|   Best |  88/300 |       92 |     45    8   39 |        1.12% |        46.309 USDT    (9.26%) | 0 days 09:00:00 | -5,028.57406 |                           
 [Epoch 300 of 300 (100%)] ||                                                                                       | [Time:  0:20:44, Elapsed Time: 0:20:44]

with pvt

+--------+---------+----------+------------------+--------------+-------------------------------+-----------------+--------------+                           
|   Best |   Epoch |   Trades |    Win Draw Loss |   Avg profit |                        Profit |    Avg duration |    Objective |
|--------+---------+----------+------------------+--------------+-------------------------------+-----------------+--------------|
| * Best |   2/300 |      139 |     45   26   68 |        0.29% |        18.105 USDT    (3.62%) | 0 days 13:37:00 | -1,301.19708 |
| * Best |  18/300 |       90 |     41   18   31 |        0.38% |        15.224 USDT    (3.04%) | 0 days 06:41:00 | -1,539.64835 |                           
|   Best |  36/300 |       99 |     39   19   41 |        0.72% |        32.016 USDT    (6.40%) | 0 days 12:29:00 | -2,799.91656 |                           
|   Best |  73/300 |       89 |     39   15   35 |        0.97% |        39.001 USDT    (7.80%) | 1 days 08:03:00 | -3,794.09449 |                           
|   Best | 259/300 |       75 |     27   24   24 |        1.46% |        49.272 USDT    (9.85%) | 0 days 10:57:00 | -3,937.80395 |                           
|   Best | 295/300 |       72 |     26   32   14 |        1.70% |        55.163 USDT   (11.03%) | 0 days 15:09:00 | -4,422.25698 |                           
|   Best | 300/300 |       65 |     27   30    8 |        1.82% |        53.369 USDT   (10.67%) | 0 days 18:15:00 | -4,921.45121 |                           
 [Epoch 300 of 300 (100%)] ||                                                                                       | [Time:  0:20:05, Elapsed Time: 0:20:05]

[nuaimat]

nit: please fix the # identation

[andif888]

fixed in 8f2b61f

[Rikj000]

Thank you for your work so far @andif888! 🦾
However please in the future let the HyperOpts finish so we can be more certain about the results 🙂
(For me it also takes hours for each HyperOpt, which might look long at first, but actually it's rather short to find things like these out in my opinion, currently my machine has been occupied all weekend HyperOpting for a good set of Freqtrade protections and I'm still not where I want them to be, only have been able to do a good handful of tests since these tests do take their time 😄)

However at the moment I can't see any clear improvements in the HyperOpt Results here can you? 👀

Also I'm thinking about only adding new indicators / weighted signals into the mix at the end of a development cycle, a little bit before packing up a release. So we'll always remain to have consistency in our weighted signals while implementing the things for the next version of MGM.

To keep track of upcoming versions and things still needed in them you can check the Milestones