Pumps/Impulse Control

You have to determine the value of the pump if you’re going to be pump chasing.

A pump by a bunch literal pump and dumpers by this Teeka faggot will end badly for you.

These people on the inside buy up and then dump on people like YOU. By the time it reaches you, you’re looking at a higher chance of being dumped on.

Couple months ago it was good to bet on it. But now with all the leaks and this PBC thing has become a bit of a joke now, it’s more risky.

This morning I was asleep during the ripple pump, but I’d be more willing to hop in that because it’s more substantial.

So maybe next time when you get that feeling of OH FUCK OH FUCK gotta get in this, maybe stop and think if it’s a good idea. The feeling of losing money is worse than the feeling of the gains you didn’t get

That is a good idea. Just pause 1 second and name the feeling I am having before I hit the buy order. I just don't trust myself to keep myself in line.

I wish I could enable a 5-second enforced delay in Bittrex where it just shows me a dialog that says "Are you fucking FOMOing? Be honest." With a cancel button available underneath.

>.5 BTC is full retard.

Jesus dude, enjoy your sick $10 gains.

Took about 4 months for me until it clicked. I think what really did it for me was when the BTC low of 3k happened, I was getting comfortable using TA at the time and noticed how it bounced from the same level the last time, so I bought at 3k, and it pretty much bounced at the exact level that I predicted. I also did a lot of dumb shit of course, and still do probably, but not getting driven by emotions when trading is something that I can recommend to everyone. But more and more I believe that just investing in solid coins and holding is better than trying to beat holding gains by trading.

>tfw still holding the nmr ATH

you retard fallow below mantra

1. if you feel fomo, close your laptop, get outdoor for 30 min
2. also check the graphs , previous ATH, also if your balls are as solid as rock, dont do swing trade. one safe bet is always try to buy below previous ATH, if miss that fuck your laptop and look for some other moon mission, in crypto daily 10 moon coins already out there, only thing picking early with out fomo.

Trade around a core position. Add on dips.

Dont go all in especially after a moon mission launches. Buy a small position that way If it dumps you can Avg down.

I few months because in my first few months I cut my btc holdings in half doing this.

I am relatively new few months back, I fomoed on xlm and ibm news, costed me 20k, heard it right. afterall that news did not get me profits as I overestimated news. learned lesson from that. now no fomo. if i feel just close laptop and go out. every time I do it, when I come back and look at it. almost all cases it helped me and feel right decision.

At the very beginning it used to happen to me over and over again, until I understood that there's always this little thing called retracement.

just added to cmc

coinmarketcap.com/currencies/faceblock/

low supply, next week binance in line.

never happens with the exception of the BCH FOMO last week. didn't loose anything because I was early, but still it could have turned out bad.
usually I just buy the dip and everything is fine

And also I learned to read indicators, I avoid buying overbought shits.

sorry wrong thrd

I learned before I got into crypto. With stocks. If you chase, you're going to get burned. Buy on weakness, not strength. Glad I learned that before getting into crypto. As a result, I don't have any losing trades.

buy high, sell _______

TA with "don't trade during news" exception.

I wrote a number of bots using various combinations of indicators to trade on Bittrex. They averaged a decent profit (0.05 btc/day) but I took them offline when BTC began its 5k+ rally since BTC-based markets went into #bearmode at that point and I had no shorting strategy ready.)

Writing bots made me understand indicators on a much better level and showed me what can be accomplished by keeping emotions out entirely.

Watching a computer executing my personal strategies with zero influence from panic/FOMO was deeply enlightening, and I recommend everyone do this if they have the desire to become great.

I now use my bots to monitor markets and look for interesting flip opportunities. As an example, I've made over $10,000 in the last week from monitoring 5min and 30min RSI on every USDT-based market and buying dips on OMG, NEO, XMR, ETC and ETH.

Heh. I learned years ago doing penny stocks back in 2005. Fuck biotechs.

But anyways... Stop chasing shit coins. Sure someone is going to rich, but unless you are releasing your own ICO, most of the people who invest in these things are going to be losers.

Example. 70% of Wallstreet active trading mutual funds do not beat the market index. Yes, that means 30% of traders beat the market, but fuck... You've got a 70% chance that you won't and will actually lose money.

Stop chasing the shit.

If you stuck with the standards (BTC, LTC, and ETH) you would have done quite well even if you got in say around $4K.

how do I become like you? did you do any formal training or did you start from 0?

This is such a good thread. Definitely saving.

what bots would you recommend?

>I've made over $10,000 in the last week from monitoring 5min and 30min RSI
Interesting. I also programmed a bot but it's really hard to find a good strategy right now. Anytime bitcoin goes crazy it pulls down all altcoins. I thought about just watching the bitcoin price and sell everything once the % price increase reaches a threshold in a small timeframe.

No formal training.

My wife showed me Veeky Forums back in May this year and I fell in love with your memes so I stuck around. Bought some BTC from a guy on localbitcoins and started playing around on Bittrex. Found Investopedia and learned some basic shit. Started watching "TheChartGuys" on jewtube to learn about TA.

After three months I was enjoying myself so much I couldn't pretend to care about my day job (software engineer) anymore so I quit.

I would recommend writing your own. I wrote 4 different ones so far, one in C++ and three in PHP.

I had that same problem and I solved it by limiting myself to USDT-based markets. They are way more stable than the BTC-based ones and while volume can be a bit lower, there's an army of cross-exchange arbitrage bots providing a sort of asynchronous liquidity here.

Cool, thanks for the suggestion. I already thought about just ignoring the BTC markets but thought that maybe, with a good strategy, it's manageable. But yeah you're probably right. I'm gonna run some backtests on USDT markets now.

I am lurking.
#4CHNallday

> Identify coin. I trade volatility 90% of the time. Volume is key.
> Grab OHLCs 1m-5m through the exchange's API as far back as possible.
> Perform basic time series tests in R.
> If trend-, seasonal-, stohastic-, etc- reduction is possible I go ahead with the coin.
> Start developing the model in R. This could be a simple linear model (ARIMA+EGARCH), could be something more complex like SVMs or trees, usually its a combination/ensemble of models.
> Usually never go as far as heavy machine learning stuff like a CNN+RNN stack. Easier to find another coin than to make sense of the non-linearity later on.
> The general idea is to take a time series and identify the generative process. In other words, reduce the residuals to white noise.
> Once model is ready, reimplement it in python as a trade signal in mementum/backtrader (on github)
> Writ the strategy. The strategy can either rely on the model for everything (long/short/close signals) or just listen for a particular signal (usually the entry point) and then close depending on some heuristic.
> Once done, run a ton of backtests.
> If backtests aren't working out (bad sharpe, over the top drawdowns) go back and rewrite stuff until it works out.
> Once its all good, either rewrite everything in Erlang or C++ (rarely needed though) or just use python and start live trading.
> Monitor the strategy (all done automatically) and retire it once its starting to get "arbitraged away"
does this sound like a reasonable approach to you? can you give us some leads to making a successful bot?

I've been trading for almost 3 years and it's a struggle keeping your emotions in check.

Yep, that sounds reasonable if somewhat gruntwork-heavy. It also sounds like you're working with more sophisticated concepts than me, so you'd be better taking advice from someone else at this point :)

Just remember that any failed trade in red will turn green with only 2 successful trades (assuming you're always putting the same amount at stake).

Then you only need 4, then 8. These are ridiculously huge chances not to take them. Sooner or later you'll realize it's almost impossible to lose. If you're right at least 12.5% of the time, you're going to make it. Just stop worrying, stick to the plan, let go of your ego burdening you with emotions where there's no place for them.

Don't let the scared animal living in you win with your rational brain.

This is an incredibly easy game, it's just hard for humans.

I don't do quant trading much but I am a machine learning grad student

This approach has a high risk of overfitting models. You also take a lot of things for granted. Doing what you describe successfully alone is pretty difficult, it's something I would be hesitant to take on as a personal task on top of my research/school even though I am nearly an expert at this point

>Simple linear model(ARIMA + EGARCH) probably as far as you want to go. There is applicability with support vector machines here but trees are useless and CNN RNN are worse than useless for this

Consider that traditional equities and derivatives are very complex and are traded by thousands of highly competitive firms. If it was just you and two other firms, the approach to positive trading would be to try to proactively identifiy their strategies through data analysis and manually design countering strategies. The market is too large to do this, so instead of trying to ever comprehend the underlying causes of market behavior from all the other firms, we throw advanced CNN/RNNs at them to try to get a symptomatic signal. In the stock market (high liqudity at least) there is relatively little noise to possible signal--stocks are traded for a reason, particularly when you can identify who's trading them, so it makes sense to train a model on every little indicator.

But in crypto, most price movements are opaque and unmotivated. You are not competing with firms who view you as their equal and are harnessing advanced tech to fuck with you. You are competing with day traders, normies buying in, and fomo from projects and ICOs, all sorts of shit that a neural net could NEVER decipher. So you are better off doing analysis + manual trading, or simple linear models, to try to understand the more macro price movements.

Simply put you're overdoing it. You'll never get past
>run a ton of backtests
because your models will always be shit at predicting crypto.

not my strategy just some copypasta. Don't understand half of the matter but i am a Veeky Forums faggot and have some ideas of autismal patterns i have picked up on from staring at these charts for hours everyday for the past year in my neetdom. I am a horrible programmer and only know python but know some serious maths.
machine learning sounds like a dream. Yes I want to utilize signals I did during manual trading and automate them. Alt relationships to BTC seem imperative. I was also thinking about having a weight system for numerous indicators to signal for buy ins on a list of reputable less shit shit coins. Retracements and the meme triangles is another scenario I want to automate. As well as catching residuals of large pumps on high volume shit coins where there usually exist one to several bounce backs. Some of these would be very conservative but I am okay with small margins as long as they are positive. Selling seems to be the most confusing and I'd be rich right now if I wasn't such a greedy fuck.