[_ll5a]

Has YC ever considered having a control group by randomly pre-accepting certain applicants into the program automatically? I'd be curious to see the stats on how successful that group would be versus the group that's actually picked by the committee.

[sama]

We do talk about that every once in awhile, but the average company that applies to YC is not very good and thus the experiment would hurt the network (and be painful for us to work with).

What we've more seriously considered is randomly funding some companies just below the threshold.

Also, we do look at our ranking to see how companies at the top vs companies near the cutoff do. It's rare (but not unheard of) for a company very near the cutoff to be one of our most successful companies.

[gogopuppygogo]

Can you elaborate on what makes a company "not very good"?

Would you ever consider releasing anonymized data on companies that apply and are accepted along with those who apply but aren't accepted into a batch? That'd give future applicants much more insight into the process.

[beachstartup]

EDIT: i should emphasize this is only my experience as a startup hiring manager. maybe YC attracts nothing but geniuses, i dunno. i'm just a guy on the internet.

since yc is obviously not going to talk shit on their own applicants, i'm going to go out on a limb here and guess that it's just like every other job application process when six figures (or more) of money is on the line. big paychecks attract plenty of, shall we say, 'aspirational' applicants.

1. apps where spelling and grammar is clearly not a priority (sure, i'll give you 5 figures a month to spell shit wrong) 2. the wrong kind of company will apply, i.e. a restaurant 3. bullshit artists gonna bullshit, even when they stink 4. people simply not smart enough to do the job. just plain dumb is a real thing, folks. 5. people with too little real world experience 6. really stupid ideas (who's to judge? well, the people with the money get to judge, that's sort of how applications work)

the fact that you can use the term 'anonymized data' in a sentence that actually makes sense is going to put you in the top 10% automatically - yes, the bar is that low when you're dealing with the general public. the trick is getting into the top 1% though. it's a power law distribution of difficulty.

every job advertisement attracts applications like these, and yc is basically just a really, really fancy and structured super-duper college/job application process combined, but for crazed entrepreneurs (like me). you are going to get a lot, maybe even mostly people on the left hand side of the normal distribution of iq, talent, and work ethic/experience. the easy thing is the dipshits and fakers are easy to spot-and-drop almost immediately. i can basically glance at a resume and disqualify it in 10-15 seconds if it's not good enough. you just can't bullshit someone far smarter, more experienced, and better at bullshitting than you. CAVEAT: there are extremely smart and experienced bullshitters out there though -- that's the real danger area. that's what you're really worried about.

basically the only thing you can count on is that they want your money!!

[newjersey]

Oh wow. I wish they said that the bar was so low. I would like to apply as well if that's the case.

Edit: I still think the idea of Dropbox is pretty stupid and I wouldn't have given Drew Houston a penny if I was an investor. But I never thought the average application would be so far worse if Dropbox got funded and succeeded.

[dang]

It's been a few years since I've reviewed applications, but yes, you should apply. Put it this way: if you simply (a) answer all the questions (b) coherently, you're well ahead.

[argonaut]

You're making the mistake of judging the idea too much and forgetting that Drew Houston went to MIT, and had worked as a tech lead at a company. That already puts him in the top 10% of applicants, regardless of his idea.

[newjersey]

I guess so. I, on the other hand went to a state school and my work experience is working on fairly straightforward line of business applications. I am not so much criticizing Dropbox as much as I'm making fun of my own lack of vision.

[beachstartup]

well, hold on, i don't speak for yc, i was just going off of my own hiring experiences and perhaps extrapolating a bit creatively.

but i can read a lot into sama's "not very good" understatement. for a yc (or vc) partner to say that in an actual post is pretty telling.

and of course, my statement about grammar is grammatically wrong. not much changes on forum banter through the decades, does it?

[orionblastar]

A lot of the not very good companies don't have a business plan, aren't organized very well, have issues, are in it for the money but lack experience and skills. They need a lot of help just to get better and I guess they don't have what it takes to get funded.

You don't want to fund a company that doesn't know what they are doing and needs a lot of help to get to that level where they do get funded.

But that is a startup opportunity to help these awful companies get better so they can get funded and then pay the mentors money for the training.

I'm sure they get coffee shops and pizza places apply as well that don't qualify but want to do some sort of cybercafe in their store.

[cookiecaper]

This is a phenomenon referred to by various terms, including the Dunning-Kruger Effect and several colloquialisms (like "you don't know what you don't know").

The short version is that there is an inverse relationship between actual and perceived skill. In general, the better someone is at something, the less they rate themselves relative to their peers.

This is a big problem when recruiting people for skilled positions. The best applicants assume that they won't be good enough and don't apply. The newest applicants have no reservations about applying and assume that they are qualified for anything, kind of like someone who graduates with a BS in CompSci and thinks it means he knows everything about being a professional programmer before he's even ever held a real programming position.

Joel Spolsky discusses this too! I can't seem to find the post right now, though. The gist is that he was encouraging good developers to apply to Fog Creek, because he noticed that a lot of developers he would have liked to hire were intimidated about applying there, as he had previously discussed things like dismissing non-amazing candidates, how rare it is for good people to be on the market, and the extensive perks his company extended to its developers (like an office with a door that shuts). The good people read these posts and automatically filtered themselves out, no doubt comparing against their idealized version of what they'd like themselves to be, emphasizing the gaps and flaws that they know are in their resume, instead of the reality of the applicant field, the competence of which most good people grossly overestimate.

There's another cheesy colloquialism that I think encompasses this well: "you're comparing your behind-the-scenes with everyone else's highlight reel". We use ourselves as the reference point for understanding others, but the intimacy we have with our own thoughts and failings can cause us to forget that they're not as big as they seem. We should be cautious before allowing ourselves to be intimidated by others, bearing in mind that we only see a small snippet of the picture.

[0] https://en.wikipedia.org/wiki/Dunning%E2%80%93Kruger_effect

[wingerlang]

What's such a bad idea about Dropbox? A single folder that automatically syncs your files across devices.

Even for less technical people it makes sense and it is easily explained. Put the files in this folder and they are safe.

[JabavuAdams]

> the easy thing is the dipshits and fakers are easy to spot-and-drop almost immediately.

You seem pretty sure about this. How would you validate that belief? What are the consequences if you're wrong?

> i can basically glance at a resume and disqualify it in 10-15 seconds if it's not good enough.

I'd suggest taking a more systematic approach. You may be rejecting some great candidates without decreasing the likelihood of false negatives.

> you just can't bullshit someone far smarter, more experienced, and better at bullshitting than you.

If you can't spot the fish, then ...

Also, why you no use capitalization? It makes your post read dumb.

[cookiecaper]

>You seem pretty sure about this. How would you validate that belief? What are the consequences if you're wrong?

The thing about saying no is that it's a lot less risky than saying yes, in terms of damage potential, not necessarily in terms of "lost upside" (which IMO is something that rarely merits consideration). If you pass on a good candidate, they will move on and find another home that works for them; no harm, no foul, so one doesn't need to feel bad about admitting that the person is skilled, but just not a good fit.

If you bring someone on board and they turn out bad, they can and often do cause substantial harm to the company/org. Once they're there, they have access to everything they need to wreak havoc, and as I'm sure you know, malicious intentions unfortunately aren't really necessary to cause problems. Big issues for your company can be caused by simple incompetence.

Joel Spolsky discusses this in his essay on interviewing [0]. It's just much safer to let the candidates that aren't enthusiastic yesses go than take a gamble and lose.

>Also, why you no use capitalization? It makes your post read dumb.

I don't really mind this, but I do think this is noteworthy since he specifically called out spelling and grammatical errors as disqualifiers. I assume in reality, he's forgiving of personal quirks or typos, and means that he rejects candidates out of hand who appear genuinely unable to spell or properly utilize much of the industry's vernacular.

[0] http://www.joelonsoftware.com/articles/GuerrillaInterviewing...

[w1ntermute]

Were the companies rejected by YC that have gone on to be the most successful often just below the cutoff?

[namenotrequired]

I recall pg previously stating that there was no correlation between how close to the cutoff an accepted candidate was and how well they ended up doing. Now Sam says the opposite. I can't find the original quote now and this was when PG was still in charge, so they probably have improved their interviewing process and collected more data since.

[tim333]

I can't answer that but pg talking about what may have been Sendgrid was quite interesting along those lines https://www.youtube.com/watch?v=YMqgiXLjvRs&feature=youtu.be...

[nikkwong]

I can second this. I worked for a similar unnamed incubator company in the Bay Area and a vast majority of the ideas sent to us were from good-hearted people, but absolutely ludicrous and—at least from our opinion—would never be able to become successful businesses. Some people just really seem to have problems understanding product-market fit.

My guess would be that applicants to Y-Combinator are probably along the same spectrum–some great ideas, some decent ideas, and a lot of junk inbetween.

[vonklaus]

If possible to ethically disclose this, what was the craziest thing you think could have worked, amd the outright craziest.

[kleer001]

If they don't answer just watch a few seasons of Shark Tank or Dragon's Den (any country). I highly doubt there's much different in business nut-bar factors. Some people just have no grasp (yet) of what it takes to get good.

[nikkwong]

Yeah. The way a person communicates tells you a lot about their ability to lead a company. Most people would pitch us ideas and we wouldn't even vaguely get a sense of what they were trying to sell—native english speakers too. Not surprising at all—but people who had chances to get funding were those who were clear & concise. Those who's pitches who we didn't understand immediately were for the most part mentally discarded and never acted upon by us.

[lancewiggs]

There are plenty around - one example where I live is a very persistent chap with some sort of mystical flying machine based on garbage physics and a track record including fraud by scamming investors.

[biot]

Or another idea along these lines might be to have a checkbox on the application that says "Submit to audience choice if rejected" and, if they don't get in, those applications are opened to voting by HN users above a certain karma threshold. Pick the five highest ranked applications and bring them in as a normal YC company.

[sushid]

Aren't you afraid that something like that can be easily gamed? Or maybe voting can be limited to accounts that are 2-3+ years old and/or Karma greater than n.

[daveguy]

I like this. An alternative would be to have borderlines ranked by those who got into YC in the current round or in past rounds. Top peer ranked would get in just under the cutoff. HN sounds even more exciting (probably couldn't include the whole application, maybe just a company description -- although that might not be something people are comfortable with). Either way, I wonder if there is any peer non-YC consideration.

[spacehome]

The fact that the voting is public would kill the blinding.

[ryandrake]

If the thought is that randomly selected start-ups might possibly outperform hand-picked ones, then I suppose anyone with a few million $$$ laying around is free to test that hypothesis. If true you could make bank!

[ec109685]

Nobody besides YC has their network. The question is whether it is their selection process or other factors that contribute to a startup's success.