Processing of external transactions for third-party trading robots
A common mistake, when an external robot is connected, but it doesn't work. The reason turns out to be commonplace, external processing is disabled…
algotrading
The robots are to blame
When the market moved up, on the empty place, everything was fine, investors happily rubbed their palms and moaned speculators-bears. But it was worth …
Advisor.
Need “buzzer” signals (with duplication in a pop-up message) to test the trading system. The idea is pretty simple and primitive (what needs to be encoded). I’ll make a reservation right away, you can't build a robot on this, entry points for making decisions and testing ideas just have to be indicated. I am preparing the terms of reference now. Trading terminals – Alpha Direct, or Quick (doesn't matter). All the details, suggestions, payment issues – in a personal.
PS I can't code myself, there is no time for training now; acquaintances are busy, but you need to do it quickly. Therefore, I turn to LJ.
Almost perfect…
… almost perfect… This “nearly” only adds piquancy. But her main advantage, that she has no stream of thoughts, emotions, opinions, and market expectations. Graceful, deadly weapon.
P.S. S&P all.
New Year's system.
Didn't write about trading this year, that way I will completely lose interest in him :)
Meanwhile, one of the tested trading algorithms, showed on the five-year history of the winratio futures 80%, with optimal take / stop settings. Others have more modest results.: 50-60%, which is also good.
The robot learns.
The counter-trend algorithm is gradually becoming overgrown with meat. New conditions for entry are being added. Making a normal counter-trend turned out to be more difficult, than trend following model, Nevertheless, on the five-year tests of the PTC index, he is in the black, bypasses buy and hold, and has a significantly lower drawdown. On the positive side: the model is universal, and can work in different market conditions without additional optimization; the amount traded can be quite decent. Of the cons: profit, for now, only slightly exceeds the yield on a good bank deposit; the drawdown is still too big. From current tasks, reduce drawdown by three times, and increase profitability, so that you can hang shoulders and capitalize interest. I figured out how to screw on another filter tonight, We will try. At this moment, the biggest difficulty, prohibit the robot from entering the countertrend in narrow flat zones, after significant movements, with the potential to continue. Sometimes he misses delicious reversals too..
Algorithmization.
В настоящий момент выработано два подхода, трендовый и контртрендовый (как оригинально, ха-ха). Трендовый имеет большее количество верных сигналов, но безопасные условия для входа возникают не так часто. Плюсы контртрендового – близкие стопы, возможность поймать разворот и взять бОльшую часть движения, но количество истинных сигналов поменьше. Для каждого подхода выработан список условий, который должен полностью сооблюдаться. Для контртренда он более объемный (с десяток основных, плюс подпункты). Когда это все собралось, естественным образом возник вопрос о целесообразности ручной торговли. Happened, что самое слабое звено в цепочке – это я сам. Ie. когда система оформилась, It revealed, что руками, в силу определенных особенностей своего характера, и образа жизни, я по ней торговать не смогу, да и не самоцель это. И довольно трудно оказалось, иногда взглянув на график, прогнать соответствие текущей ситуации всем факторам, на различных инструментах. So that, вполне логично появились предпосылки к роботизации. Хотя сам я, long time, был противником МТС, в проекции на себя.