s it reliable to use data algorithms to evaluate venture capital projectsThe nternet Wangzhuan 3 y

contact Wangzhuan is 2007, a colleague is engage in Wangzhuan search by foreign Master, make a pot of gold, it is said that he was a month can reach around $1000. He introduced me to understand the next mail click, so in 2-3 months of foreign mail click, a donkey standing very famous at that time, I had to manually, and then find a software to the point, the end result there are a few dollars? Don’t clear. But because there was no PP will be confiscated. By mail Click to search later like this one, but colleagues said that the search is the decline of the industry, and then do the search page trouble, under such circumstances, the Wangzhuan project ended. Let me begin to recognize the mail click Wangzhuan, this new industry.

artificial intelligence in the financial industry, there are many real value can be applied scenarios, in which the two market investment transactions, data based quantitative trading algorithm has also been a lot of successful operational case. The investment community’s interest in data analysis is increasing, and in the primary market, venture investment, and even angel investment, there are some attempts to assess the value of project investment by analyzing data.

we are worried that the machine has ruled the people, while sparing no effort to push all the responsibility to the machine, physical or mental is also. Venture capital investment can not rely on data algorithms, there are many paradoxes.

so-called venture capital, the investment community of Pyramid, especially angel investment, is the highest risk category in the investment field. The reason why capital is willing to take a high risk is that such capital is chasing ultra high returns. The higher the degree of risk to the corresponding high yield can be expected, this is the nature of capital. The most conservative investment allows almost all investors to get the most guaranteed income.

now calls for constant precision based on big data through machine learning techniques

2008 began to engage in Wangzhuan forum, the forum is often the Wangzhuan home, so want to engage in a home network to imitate their Wangzhuan forum. The forum program is discuz, this program is very excellent, very suitable for novice webmaster. Then dedecms would like to use the forum expanded into a webmaster information platform, but because the navigation interface change is not good, the information platform from the back to the forum category. At that time his post in the forum every day, hope to post on forum, colleagues told me now is the lack of forum posts. Also engage in some modifications on the discuz template, is indeed very laborious. With the passage of time, I found someone registered forum and post, was very pleased, but later found the advertisement is irrigation machine. At that time to engage in the promotion and chain only a little fuzzy concept. Then apply alimama ads, they will often remember a few, have an income of 2 yuan. Due to space, also stopped on the business forum. The forum has allowed me to deepen the understanding of wangzhuan. I never think Wangzhuan is a derogatory term, at least is a neutral term.

in fact, it’s the same thing to vote 1000 times and 500 times 1 times, because you probably won’t have more than 500 investment opportunities 1 times.

paradox 2: algorithms need data to support, while VCs data are superficial

paradox 1: big data algorithms are good at finding probability events, and ultra high returns must be small probabilities of

Abstract: venture capital investment cannot rely on data algorithms. There are many paradoxes here.

, for example. The likelihood of A success is 80%, the likelihood of B is 90%, and B is 12.5% higher than A, and it’s easier for you to decide to bet on B. This is a comparison of the probability events.

, only a small number of investors are able to achieve very high returns in high-risk investment categories. As a result, we can understand that the venture capital is bound to find a small probability event, the so-called unicorn, the cream of the crop. The algorithm based on large data is universal, that is, its ability to look for probability events. The underlying intention of the two is almost the opposite.

, and if A and B are both small probability events, things will get a lot more complicated. For example, the likelihood of A is 0.1%, the possibility of B is 0.2%, it looks like the B success rate is two times that of A, but at the same time, this also means that the possibility of A probability will not occur is 99.9%, B is not 99.8% possibility, the possibility they are not nearly the same, and have a high probability. On the basis of such data will still hesitate.


also started in 2009 to sell SOCKS5 agent, the first not to sell later had some improvement, but then the supplier to stop delivery. The fundamental reason, the market only a home network, S5 is much better than other agents of their home, so they decided to monopoly profit more. There is a direct cause of people in the home below their pricing sales, so they put all of the agents are blocked. But to sell the S5 proxy, >

in 2009 by online friends in ppc1000, he gave me the SOCKS5 proxy and American information, and then use the EMU method to register advertising, is the kind of registered toil. At that time he gave me serious S5 agent quality is poor, slow death, spent about 4 hours to register, after a month of $5 and hit the inside of my PP, but also with the $5 above the Yahoo bought two domain names. Use the EMU way to make money, some people can earn a lot of money, but this way is not suitable for me.

The starting time of

Leave a Reply

Your email address will not be published. Required fields are marked *