Is this really the right approach? I remember PG writing that if YC were truly rational they should have a bigger proportion of failures (because they should be taking much "riskier" ideas, because of the power-law distribution of outcomes).
Since YCombinator depends on finding Dropbox and AirBnB scale startups, I imagine one of the most important metrics to them is whether they've rejected a business that went on to achieve that scale.
Of course being in YCombinator can increase a venture's chance of hitting that type of success, so keeping track of who they rejected is just a proxy for how well their applications strategy is. They may have rejected a company that _could have become_ AirBnB or Dropbox. But this is less important since in this case that would indicate that this ability lies with YC, not the startup, and so it's less important which companies they pick. If true, this would run counter to the founder-centric model they profess, so I'm skeptical it's how they feel.
This question of who to let in reminds me of the tradeoff of precision and recall in IR. A large class can reasonably be termed higher recall, while a class where they actively exclude companies with predictors of failure increases the precision. THe two metrics are inversely related, but it's possible to have high precision and recall if you're only looking for a very few number of things and you pick them all.
It seems to me that culling out likely failures would only make sense if the partners had determined that the presence of these likely-to-fail companies negatively impacted their ability to help other companies reach their potential. This seems a likely rationale to me based on PG's comments about how dying startups take up so much of their time.
So what it really seems to represent is a vote of confidence in the partners' ability to help startups increase their chance of massive growth. And/or a recognition that likely-to-fail startups have deleterious effects on the rest of their batch that outweigh the likelihood that they'll be successful outliers. Either way it doesn't seem to have much impact on companies that apply.
It will be interesting to see if PG writes a How Not to Apply essay. Such an essay might destroy the predictive value of these behaviors, but if it stops the behaviors and they were causally linked to bad outcomes then it's a net gain.
I think your definition of "bad" implies a high likelihood of failure. Whereas I'd say that a "bad" startup is one with a low growth or profit potential; my guess is that pg's definition is similar. A bad startup can "succeed" and a great startup can fail -- they're totally unrelated things, IMO.
Here's hoping pg writes an essay on what YC thinks the "predictors of failure" are. I'm sure some applicants for S13 are probably panicking, having optimized their application for the "old" version of the YC selection process. I'm not too worried, as we're exceptionally strong on the one data point he's provided.
Much like the Google guidelines on SEO, I would imagine pg's response to this would be something like 'the best way to optimize your application is to ignore optimizing your application' and instead focus on building something awesome.
He's written quite a bit [0] on what they look for in founders. I'm not talking about pg writing a "How to optimize your app" essay, but literally "Indicators of Failure"
Right, you want him to give you more signals about what they are looking for so you can tell them what you think they want to hear - that's the wrong approach. You should instead focus on being awesome and hope that YC recognizes how awesome you are.
The perfect application may look identical to the perfectly gamed application - but the intent is different. And it's a bit naive to think YC isn't really, really good at sniffing out that intent.
> And it's a bit naive to think YC isn't really, really good at sniffing out that intent.
At this point, YC probably has a useful amount of data. I wonder if he's thrown a Bayesian filter at it? One of the things PG noted in his "A Plan for Spam" essay, is that his Bayes filter flagged indicators that he never would have thought of. I wonder what kind of data someone could get from a corpus consisting of YC's data, plus web searches on the applicants?
> “There are hardly any startups in this batch that are bad,” Graham said.
But the following [1] says:
> Y Combinator founder Paul Graham said this he feels like the contracted size of this class means that there are essentially no weak startups in the bunch
There's a huge difference between "hardly any bad startups" and "essentially no weak startups" so
it seems someone at techcrunch is misquoting pg.
Hey, I'm the writer who wrote the second example cited here -- very good point in showing the difference between the two, so I wanted to try and clarify.
When Graham said the "hardly any" line on stage, it drew a pretty big laugh from the audience -- it seemed like a bit of humor on his part. The earnest meaning and connotation I took from that (and comments Graham gave off-stage) was that there were no obvious weak links in this batch.
Thank you for the additional insight and context for the quote,
and your paraphrase. The quote makes a lot more sense as a joke,
so it would be good to note that in the article.
And if there is a bad startup -- like one that looks like it's going to run away with customers' money (or at least make them panic about it) -- don't worry, they just clip out any mentions of the YC association.
Wakemate would probably be a better example of a YC disaster. Didn't quit run away with its customers' money, but had some serious PR and internal problems...
It doesn't bother you that "YC 'XX" appears on every article about a nascent company regarding good news, then gits stripped from the title when one of them appears to be dicking over its customers?
No, it does not. The (YCwhatever) doesn't even get slapped on all stories about YC companies. There is a rather large difference between the ones it goes on and an attempt to contact the company through HN, and I don't see any evidence that the sole differentiator is positivity.
But even if it were, not including it in the title of negative articles on HN would hardly be the same thing as removing any mentions of their association with YC.
"He [pg] said that this time around, YC looked at 'predictors of failure,' not just 'predictors of success.' For example, he said that in the past YC might have chosen a company that had great founders (a predictor of success), but this time it might have filtered that same company out because those founders, while great individually, all hate each other (a predictor of failure)."
The given example implies that YC has chosen founders who don't like each other in the past. Seems remarkably odd and hard to believe IMO. It may have meant to say that the founders seemed LIKELY TO hate each other eventually but if they even disliked each other at the interview, I can't imagine the YC partners being ok with this.
I think he was just giving an easy to understand example which didn't really reflect reality. The problem with predictors of failure, unlike success, is that it's a lot easier to hide them once you know what they are. If you know, for instance, that having founders who when asked say they're doing a startup because they hate having a boss, no non-idiot will give that as an answer when asked, even if they believe it, ruining the predictor.
It could simply mean that they never spent a lot of time investigating how the founders got along, and so weren't able to notice initially which founders ultimately didn't work together well.
Why the big fuss about saying some startups are "bad"? I'm sure some startups are bad. There's certainly a possibility that some bad startups looked good and made it through the screening process. Other accepted bad startups may have been risky bets that didn't payoff.
I think at least a few people have a problem with the directness and harshness of the statement. But if I was one of those bad startups, I'd appreciate the candor and directness. That would be far preferable to someone being "nice" and allowing me to continue down the wrong path.
This is opposed to last sessions where Paul Graham said lots of the startups that just started and he had subsequently just allowed into his program were bad?
I know Y Combinator is selective. However, claiming a 0% false-positive rate is just plain denial. Or possibly a redefinition of "bad" start-up or "bad" founders.
It comes across as arrogance, and "pride comes before a fall" and all that. I would also definitely say that the Valley hype machine has been producing what sounds an awful lot like its famous irrational exuberance lately, yet meanwhile VC returns-on-investment are not actually that good.
However, claiming a 0% false-positive rate is just plain denial.
But that is not what he claimed. He described what he meant, and what he described says that he could see about 90% of the startups being in sufficiently good shape that they could still win big. That means that he thinks about 10% are definitely failing right now.
I believe one of two things about this statement, neither of which is what pg likely meant by it:
1. YC has bad startups it just hasn't recognized yet.
2. YC has an entirely mediocre batch of startups due to not taking any risks on "bad" startups.