5 ms·

Here in New York City & New Jersey, bail reform was passed circa 2019. This limited when bail could be imposed on people, including the almost complete eliminat

Here in New York City & New Jersey, bail reform was passed circa 2019. This limited when bail could be imposed on people, including the almost complete elimination of mandatory bail. It is completely up to the District Attorney's discretion. The idea was that illegal drug dealing shouldn't come with an unusual burden for bail, given that its non-violent. This was heralded as a victory almost immeadiately: there was little change in crime rate within the first months of it being passed. This seemed really strange to me. Why would you declare victory like this so prematurely? It was just _months_ after it was passed that people said "look no difference, more humane." Such proclamations should be waited at least two years so you could handle any noise. Also noted in this thread has been the new hyper progressive DAs who don't prosecute unless its really bad. Take SF with Chesa Boudin, who doesn't prosecute thefts under the felony limit (roughly $1000) _at all_. There's a lot to be said about whether these all form moral hazards but I leave that to one's own interpretation. I will say that people under ~35 never knew a New York City so dangerous you couldn't walk around at night.

BoorishBears 5y ago
Why two years? Why do the people interacting with the system on a daily basis not get to make observations on a major change in the system in "just" months?
- bingohbangoh 5y ago
  You can't say "there was no difference in crime" with only a few months of data. It's too noisy. There isn't enough data.
  - lesuorac 5y ago
    > It's too noisy. [Citation Needed]
    - bingohbangoh 5y ago
      I'm surprised two people are debating this. Statistics over small timescales, e.g. months, will not be as accurate as statistics over large timescales. This isn't disputed (unless it is? I'd be curious if you know people who think statistics with fewer numbers are more accurate than statistics with more numbers. As for if a few months of data was enough to proclaim victory, NYC's current surge in crime seems evident to me that a few months was not enough data. As for if this is directly relevant to the surge in crime, I leave that to your own conclusions. As for why I picked two years, it was an arbitrary guess. My point is that we need more data before announcing "bail reform solved problems."
      - redis_mlc 5y ago
        > I'm surprised two people are debating this. With Marxists, everything is a debate until their side wins. As we're seeing with the murder rate, Afghanistan and the southern border, that means we all lose as a country.
      - BoorishBears 5y ago
        You're being insultingly disingenuous. No one argued that statistics over a larger time will not give more signal. We all know what basic stats is. The question is why you get to override the opinion of the people directly involved on what a suitable period of time is. It's farcical to imply that a period of time defined by the people who are literally on the ground for this is somehow unacceptable, yet the unit of time you literally pulled out of your behind (oh sorry, "arbitrarily guessed") is somehow acceptable. - "Small timescales" is relative to what is being measured in case you didn't know... for some observations 30 seconds is an eternity... So what about we leave determining what a suitable timescale to make observations on is to the people who aren't making arbitrary guesses from their armchairs and were literally watching this unfold in their day to day lives.
        larksimian 5y ago
        Given the national nature of the increase, it's unlikely that a local measure caused it, it's probably pandemic-times related. That being said obviously a longer time scale is better for trend detection than a shorter one.... Oooor is it dundundun :D since if you pick too long a timescale then you get additional events muddying the causality, like aforementioned pandemic. That that that being said public policy changes are chaotic things that can have one effect in the short run a different one in the medium term etc. It's kinda meaningless to declare victory or defeat, as society's interaction with laws evolves over time.

BoorishBears 5y ago

Why two years? Why do the people interacting with the system on a daily basis not get to make observations on a major change in the system in "just" months?

bingohbangoh 5y ago
You can't say "there was no difference in crime" with only a few months of data. It's too noisy. There isn't enough data.
- lesuorac 5y ago
  > It's too noisy. [Citation Needed]
  - bingohbangoh 5y ago
    I'm surprised two people are debating this. Statistics over small timescales, e.g. months, will not be as accurate as statistics over large timescales. This isn't disputed (unless it is? I'd be curious if you know people who think statistics with fewer numbers are more accurate than statistics with more numbers. As for if a few months of data was enough to proclaim victory, NYC's current surge in crime seems evident to me that a few months was not enough data. As for if this is directly relevant to the surge in crime, I leave that to your own conclusions. As for why I picked two years, it was an arbitrary guess. My point is that we need more data before announcing "bail reform solved problems."
    - redis_mlc 5y ago
      > I'm surprised two people are debating this. With Marxists, everything is a debate until their side wins. As we're seeing with the murder rate, Afghanistan and the southern border, that means we all lose as a country.
    - BoorishBears 5y ago
      You're being insultingly disingenuous. No one argued that statistics over a larger time will not give more signal. We all know what basic stats is. The question is why you get to override the opinion of the people directly involved on what a suitable period of time is. It's farcical to imply that a period of time defined by the people who are literally on the ground for this is somehow unacceptable, yet the unit of time you literally pulled out of your behind (oh sorry, "arbitrarily guessed") is somehow acceptable. - "Small timescales" is relative to what is being measured in case you didn't know... for some observations 30 seconds is an eternity... So what about we leave determining what a suitable timescale to make observations on is to the people who aren't making arbitrary guesses from their armchairs and were literally watching this unfold in their day to day lives.
      - larksimian 5y ago
        Given the national nature of the increase, it's unlikely that a local measure caused it, it's probably pandemic-times related. That being said obviously a longer time scale is better for trend detection than a shorter one.... Oooor is it dundundun :D since if you pick too long a timescale then you get additional events muddying the causality, like aforementioned pandemic. That that that being said public policy changes are chaotic things that can have one effect in the short run a different one in the medium term etc. It's kinda meaningless to declare victory or defeat, as society's interaction with laws evolves over time.