a pic how long is performing with stoploss is 0.5%.  Short had only 2 trades and around breakeven.

I was experimenting with pivots a month ago on this code. If compared published v130 to v135 for long only, the old version has better performance. Not worth it to put this online basically. Here’s a comparison.

Hi there – I’m running DJI code (DJ-3m-Vectorial-V3.itf) on a live account (testing it for now). Great job! This is by far the only code I can trust most. (: Tonight, I think, I may have found the first bug in the code. Running the code as both indicator, and the job on DJI 3m TF: The code entered a long trade at 25,974.7 on 8:30:01 P.M (NZ time). The trade was kept open until 10:15:01 P.M. when the market reversed into bearish trend, and the code did the right thing to sell the long position at 25,859.1, and opened a short one at 25860.1 at the same time. Meanwhile – watching the indicator (with the same code as the job) on DJI 3m TF – there seems be an inconsistency between the performance of the AutoTrading job and what the indicator shows. Specifically – closing down the long position, which unlike the job’s action, should have been closed much closer to the reverse point of the market according to the indicator. (at 25,971.9 on 9:36 P.M.) Any thought on why the output behavior differs between the job and what the indicator reports? I would appreciate any of the contributor’s feedback with this. Thanks to each and every one of you for pushing this code forward. – Rikki