Capital needed for algorithmic trading at the retail level

[JSOP]

Actually, I am a grad student, so I am free all day everyday except for my supervisor's weekly group meetings.

I would like to write software for intraday trading algorithms, but I am not sure why I won't have to build an automated trading system.

What are you looking to do exactly? HFT or just trading in general but via automation. Both the trading algorithms whether intraday or daily or for longer periods and automated trading system are of the same. They are both trading automatically via machine codes but HFT is more of trading mechanism and automated trading system is more a structured set of trading rules.

But if you are looking to do HFT, forget it, $20K is NOT nearly enough. If it's trading in general, for some instruments, $20K is enough.

 

[birzos]

How much capital is needed for someone who is new to algorithmic trading (but has the expertise to develop his own trading infrastructure) to begin trading?

Is 20,000 dollars a far too small amount to begin trading?

It makes no difference, the figures are the same.

The reference for most traders in the top 1% (the ones who make profits) is 1-2% return per month on capital using Hourly timeframes. The smaller the timeframe the more trade opportunities available, but it also becomes exponentially more complex along with associated infrastructure costs. These are institutional grade high probability trades, retail trades in the whipsaws so you can expect on Hourly one trade per day, professional will be one step removed at one per week, but obviously with more losses and higher trade volume.

That is how it works, as your day job you will need $100,000s capital, as a secondary (20%) income you're around the correct starting figure on Hourly timeframes.

 

[shatteredx]

How much capital is needed for someone who is new to algorithmic trading (but has the expertise to develop his own trading infrastructure) to begin trading?

Is 20,000 dollars a far too small amount to begin trading?

What are you gonna trade? Stocks? Futures? FX?

You should probably be able to calculate this based on your ATS. Personally, I calculate it using a combination of historical maximum drawdown and the exchange's current margin requirements (I'm trading futures).

 

[murray t turtle]

What are you looking to do exactly? HFT or just trading in general but via automation. Both the trading algorithms whether intraday or daily or for longer periods and automated trading system are of the same. They are both trading automatically via machine codes but HFT is more of trading mechanism and automated trading system is more a structured set of trading rules.

But if you are looking to do HFT, forget it, $20K is NOT nearly enough. If it's trading in general, for some instruments, $20K is enough.

%%
Most likely true, JSOP, starting capital.;
most likely most do not learn to trade with $20,000
Algo is not the key to FtKnox, ask Knight capital one of the largest, acquired by GetGo, or Getco....................Nmarish learn to trade also; good question, most all automate a 200 day moving average==============================================

 

[digitalnomad]

The reference for most traders in the top 1% (the ones who make profits) is 1-2% return per month on capital using Hourly timeframes. The smaller the timeframe the more trade opportunities available, but it also becomes exponentially more complex along with associated infrastructure costs.

This pretty much sums it up. Without a true understanding of signal processing, don't dare jump into the whipsaw. Donate your money to charity instead.

 

[Stratos Capital]

I guess it depends on what kind of instruments you will be trading. For futures, 20k capital means you won't be able to trade some markets or size your positions properly for good risk management.
I think the most important part is your position size. If you take big trades, its suicide.

 

[birzos]

If you take big trades, its suicide.

What is always fascinating with these types of question is that people don't understand the simple dynamics of trading, 1-2% per month is the target range on hourly, 3-5% if you are very experienced, what you are alluding to is the 7-10% per month range.

In the end it makes no difference which instruments you trade, if you over leverage your returns for your level of experience for the timeframe you are trading you will fail. If you try and generate a 7-10% per month return on gold micro futures you will implode exponentially faster than a 1-2% return on ES, it is embedded in the markets, the asset class becomes secondary.

 

[alex314159]

Two considerations:
1/Instruments you'll trade. Take equity index futures. CFD you often get raped on bid/offer which limits the type of strategies you can use. But on exchange one S&P e-mini contract is worth ~$200k and you usually have to deposit at least $10k to trade it. So in general you can't trade that with a $20k account. FX would be easier.
2/Target return. Developing algo trading infrastructure, if you want it fully automated, is hard. Away from the intellectual challenge, even if you could do 20% a year (very, very respectable), that's $4k a year on a $20k account. Is it worth it?

 

[fordewind]

How much capital is needed for someone who is new to algorithmic trading (but has the expertise to develop his own trading infrastructure) to begin trading?

Is 20,000 dollars a far too small amount to begin trading?

If your worst DD is $600 bucks, then you can start with $20K.

 

[TAnalyzer]

Maybe i missed it somewhere but what are you planing to trade? futures? commodities? forex?