I've made a shitty php bot for bitfinex. it buys if price goes up and sells if price goes down. with 3...

>checking the price each minute
i'm too fast for buying high selling low. if it's a rally then i ride it, if it inverts then i close and lose 0.5%

That's actually genius, it's like a brute force bot.
Reminds me of PSAR people who put stop losses at -0.1%.

>it buys if price goes up and sells if price goes down
Truly Veeky Forums incarnate, good job user.

trust me, i'm sure it will work.

i'm still fixing shorts, bitfinex' api documention is awful

So when will it be up and running?

very soon. longs work well, and closing orders too.

the decisional logic is perfect.

the only problem is that short requests aren't working.

anyway, maximum 30 minutes and it's running

My experience with my bot has been that the most advantageous times to get in are when shit has flash crashed. I know they say not to catch a falling knife, but it's been working well for me so far. +30% in a month and climbing (even despite all this fucking fork drama)

What's the decisional logic? Or would you rather keep it a secret?

no PHP is total garbage. It's the official language of rural India

Node. It's what Python always wanted to be, and much better suited for web.

good idea for an entry, but it's too bothersome to code.

checking previous price and setting a bunch of ifs is way easier

i wrote it in the op. check price each minute and ride the wave or cut losses.

>python
>weakly typed

Aren't minutes a bit too slow? Why not follow the buys/sells in the realtime and weight them by volume?

it's exactly 1 minute, which is good enough since it should avoid the retarded random dips and shit. maybe 30 seconds would be better, i'll check later

for example, in the last 5 minutes of testing, the bot bought at 1306 and sold now at 1238

Buy high sell low, the official biz bot

i ment the opposite, ffs

Please dont tell me you didnt even backtest this "strategy"

kek

>implying

oh, and this is the last 24 hours

You releasing this, or do we have to implement it ourselves form your one line description?

the way your strategy works is buying for e.g. 10 minutes each minute, so 10 times at an increasing price each time
then, if at the 11th minute the price is less than at the 10th minute, you sell all previously bought btc
so you sell 9 trades for profit, the 10th is at a loss.

how do you save yourself from trading fees?
i don't know how high bitfinex' fees are, but i'd take a guess with 0.25%

i'd imagine that the high buy frequency works perfectly in theory, but the price has not enough time to increase so far that your fees are covered

the same strategy would work almost perfectly during 2017's summer if you took some hours as your base time unit

am i making sense or do you not get what i'm talking about?

looks like the short are working, almost done.
you can post this "code" into pine editor on tradingview to backtest it: //@version=2
strategy(title = "Poorfag's road to Lamboland", default_qty_type = strategy.percent_of_equity, default_qty_value = 100, initial_capital = 10000, overlay = false)

buySignal = ( close > close[1] ? true : false)
sellSignal = ( close < close[1] ? true : false)

if ( time>timestamp(2017, 11, 11, 0, 0))
strategy.entry("simpleBuy", strategy.long, when = (buySignal == true))
strategy.entry("simpleSell", strategy.short, when = (sellSignal == true))

if ( time>timestamp(2017, 11, 11, 0, 0))
strategy.exit("simpleBuy", "simpleBuy", when = (sellSignal == true))
strategy.exit("simpleSell", "simpleSell", when = (buySignal == true))
it's more stupid. i just buy once the price start rising, and sell/short once it starts falling. pyramiding is too hard to code atm

>how do you save yourself from trading fees?
usually i don't, but bch goes +-10% in minutes, that's how

What site/interface is this?

I like the simple logic. No ai and ml buzzmemes

I'm not super familiar with Pinecode, what is the "timestamp"?

tradingview

anyway, looks like the bot is running. let's see if i coded it properly

since backtesting only supports 2k trades, it's used to limit the first trade to 11/11. technically it's not needed

Ah. So it's pretty much "last candle closed over the previous candle" or "last candle closed below the previous candle". Easiest thing ever, amazing.

Obviously you do some price tracking after that to check for the inversions, right?

Do you think checking against simple uptrend/downtrend would help the bot or only make it unnecessarily harder for it?

I'm a huge fan of your solution, it's simple as hell and easy to implement.

you can make it 100x time more efficient but i'm lazy, i just want to get the barebones version running atm

Can I have the download link? i just lost a fuck ton on bch

Be careful now when people are dumping BCH for the difficulty adjustment

the bot can short too, no worries.

i'm moving everything from my rpi to my desktop so i can run it without a hassle, 10 more minutes

I want to buy it
Can you sell it to me

If you don't know how to implement it yourself, you probably couldn't get it working.
Get some API keys first.

if it works yes, i'll sell it for pics of sharpies in poopers

for backtesting yes. the php code it 100 times longer

So it doesn't work yet? Fuck off

rude. there isn't that many algo threads on Veeky Forums so don't hate him for trying.

i'm working on a python script! keep up the good fight user-sama, php and you can do it!

wifi is shit, i'm moving it to another pc. and i need to install php here

user I'm sorry
Add me on @bendthatdick telegram
I know a some Java lel

>So it doesn't work yet? Fuck off
Read the thread, dumbass.

I just implemented this and backtested it with minute candles I've procured since 2011 starting with $10,000 and it lost every penny. I also tried not starting the strategy until 2015 and it still lost every penny.

it only works now on bch because the volatility
and volume make it possible to cover the exchenge fees

Can you show your work?

What's a pine editor?

Buy/sell price is just market close of current row
Calling buy() or sell() twice consecutively is equal to just staying in the current strategy (it checks).

Oh and btw, I changed from comparing closing prices to comparing SMA15, and it actually tripled the starting money on $10,000 to my surprise lol.

Formula for buy/sell:
tradeFee = 0.0025
feeCoeff = 1 - tradeFee
USD = price * BTC * feeCoeff

startDate was Jan 1, 2016

readRow(row) {

price = row.close;

if (lastRow && row.openTime > startDate) {
var buySig = row.SMA15 > lastRow.SMA15,
sellSig = row.SMA15 < lastRow.SMA15;

if (buySig) buy();
else if (sellSig) sell();
}
lastRow = row;
}

it's a an editor for trading view's coding language

oh, and looks like i can't install php on windows

You can, use xamp.
If you're running it from a command line, there's gygnus. Or just install vagrant and do it that way.
chocolaty would also work if there's any kind of windows package.

so a 24 hour backtest? lol sounds pretty reliable. Why dont you run one over lets say a months data, with fees and slippage. Then post your sharpe and your drawdown?

i'll run it on the live data once xampp finishes downloading

why tho? you could easily test it on historic data without risking any money

i tested it on the last 2 day on trading view. as i said, this shitty strategy can only work because of the current volume/volatility of bch. so it's now or never my friend

You don't need to play with real money to test live. It's called paper trading.

oh I see, missed the trading view bit. Assumed you were about to run it live as in real money, real exchange. Good luck though anons, I'm trading the bch/btc volatility myself but with a hedging model

Uhh guys...

10K to 1.5M in 4 years?

update: short and longs are working, i've setup everything on windows. now it doesn't close orders... but it should be easy to fix.

almost done famalams

Keep up the effort. Sounds rather interesting

yeah that looks pretty wrong lol...

>rap shows
>not beneficial to an extroverted personality type
I recommend all the rap shows

Fee and spread alone gonna kill your profit

right, should be a lot more

Did you back-tested the algorithm OP ?
Also, where can I get 1min candles
data to train algorithms ?

stay poor pleb

yes, read the thread

almost everything is ready btw, fixed closing positions

coding an algo right now to try and exploit your algo thxfam

seriously though that comment i replied too was dumb af. go to a rap show and there's about 1000 people trying to fit in and being goofy

Can't wait til you lose all your money.

good job, but ...why the fuck would you choose .php?
>of all of the hands that you might place you digits into, you place them in the hands of the people who will not hesitate to put your digits into their pockets

damn thats my female form/wife

fug

>exploit

more people doing this shit the more it becomes a self-fullfilling profecy. i'll share it for free once it's foolproof

the first btifinex api example i found was in php. too late anyway

it's on!

already made 22 cents including fee

nah you didnt get me bro, 'exploit' as in 'exploit your strategy'. I'm only kidding though, aint got
not time for this shit. Good luck with your script though I'll be watching this thread

Great. And it only works for BCC? Could you set it up for other currencies on bitfinex?

fair enough bitbro
>good fortune to you
>but the idea needs to be converted(later) to a language that isn't compromised by a people who are born spies, deceivers, liars, manipulators and thieves at heart. no offense

like which one? i can backtest on trading view but i doubt it'll work since the volatility isn't this high

Litcoin and Monero? I get that volitilty isn't high but you want a bot that can work even when volitilty is relatively low. Right?

topkek

yes, but not with this shitty strategy. if you have a proper strategy for those coins then altering the bot is easy

anyway, since the price went up and i can't affor 0.02 bch (minimum trading amount), i'm moving some eth to the bot's fund.

it will be back up once the transaction is complete and i'll post live updates

>being emotionally invested in a programming language instead of just seeing them as a tool
>not realizing that the best programming language is the one you're most familiar with

i just tried.

for reference, i got +15000% on bch in 24 hours.

i get +78% on ltc and +95% on xmr

counting fees and slippage, it doesn't look that profitable on those

Do you even understand what his criticism is?

have you tried changing the logic in the code to take fees into account? e.g. subtract fees from the sell price and add them to the buy price.

where did you get the historical data?
i've been meaning to write a bot to flip ETH or BCC so that i dont have to put any effort into being a whale
i know you can get from e.g. bittrex. com/Api/v2.0/pub/market/GetTicks?marketName=BTC-BCC&tickInterval=oneMin
but it only goes back 10 days

>being a whale
becoming a whale, i mean

trading view's pine editor has all the data you want

OP can you pls share your resources I want to try and make one, I know python but I have no idea where to begin with this

>have you tried changing the logic in the code to take fees into account?

i tried it on trading view, with the taker's fee i got "only" +60% in 1 day. with maker's fee it's +1500%.

i think there is an option to do post only order in the api documentation

thanks, is it downloadable though?

google [you favourite exchange] [your favourite language] bot and tweak it

doubt it. exchanges usually makes you pay for downloading a lot of old data

thats a shame, i'll see what i can do about it then, maybe 10 days of 1-minute interval data is enough
hopefully i'll get something working so i can brag about being NTR'd by my own bot on Veeky Forums

>i think there is an option to do post only order in the api documentation
Is that even possible? I know kraken determines it based on whoever placed their order later paying more.

Will you be sharing this code with us?

oh, found it. i'm placing order's with post only and close them without

according to tradingview that should be +300% or something in a day

if it works yes, but it's a shitty strategy just for giggles. it's way harder setting up php server on windows than writing the strategy

How much are you putting into it to get these numbers? I assume the simulation on tradingview has a limited amount of one currency at the start.