3 ms·

> This being a perfect example. The former does not imply the latter. Nor would the latter imply that there wasn't foul play. It would imply literally nothing

> This being a perfect example. The former does not imply the latter. Nor would the latter imply that there wasn't foul play. It would imply literally nothing of interest and is an absurd claim to lead with. Congress beat the market absolutely doesn’t follow from 1/3 of them beating the market. It would be strange if that number were considerably lower. > First off, there is a third option: matching the SPY. realistically only possible if there’s a healthy margin to be qualified as “beating” it. Is that considered here? Idk. I’m guessing not. > Second, beating the SPY is a proxy metric which is typically used to help identify those that have performed suspiciously well, not those who performed average. It is actually difficult to beat the market, and more so to do it consistently[0]. So I must insist that you use the metric as a guide, not a filter or target. No, that’s a stupid claim. It’s incredibly easy to beat the market in a single year by chance. Again for unqualified definitions of “beating it”. It is extremely likely that a sizeable proportion of any investing group will beat the market in a given year. You are correct about multiple successive years. > Third: There are many ways to under-perform compared to the market, including not investing. I mention that specific mechanism as one thing the report shows is that there are far fewer filings by Republicans so this is necessary context to properly evaluate some fact like "Democrats beat their Republican colleagues by a massive margin." (Dems ~= 2x Reps in transactions) We must also consider that democrats have fewer members and that >50% of each party's transactions originate from a single member of that party (Ro Khanna (D) and Michael McCaul (R)). There are also multiple vehicles for wealth generation so while stocks should be of interest, it shouldn't be used as the only evidence of manipulation and/or foul play. Incorrect. The metric being discussed only included trading members. > Fourth: consider what information is marginalized by the metrics. Pareto distribution should be expected, and it is even observed. You must therefore be considerate of aggregates as variance is high (difference between mean and median is large). I don’t care man. If less than a third are doing better than the market then there’s not a lot of evidence to suggest they’re using info to beat the market. Because they’re not beating the market on average. The total return was within a few tenths of a percent of the SPY index. It’s a dumb claim. Is there insider trading? Maybe. Idk. But regardless. Do they on average beat the market? That is definitely not supported by the data. Fee free to run a t test.

godelski 3y ago
> Congress beat the market absolutely doesn’t follow from 1/3 of them beating the market. It would be strange if that number were considerably lower. This is why I stressed considering what is marginalized out. I'll use an extreme example. Let A = [1,1,1,1,50000] and B = [100,100,100,100,100]. Is A > B? Mean(A) = 10k, Mean(B) = 100. Median(A) = 1, Median(B) = 100. Similarly Congress can, on average, beat the market while 2/3rds of them do not. I hope we see that the math even allows for us to approach 0% of members beating the market while an average value can. I am guessing the reason you disagree is likely one of two false assumptions: you've confused mean and median, or you've assumed performance is reasonably homogeneous (i.e. low variance). > No, that’s a stupid claim. It’s incredibly easy to beat the market in a single year by chance. I am certain you know that "difficulty" is about intentional skill and not a reference to an act of god. You have removed any doubt that a conversation can happen in good faith. > I don’t care man. Clearly you do. But I actually don't. I don't wrestle with pigs.
- jncfhnb 3y ago
  > This is why I stressed considering what is marginalized out. I'll use an extreme example. Let A = [1,1,1,1,50000] and B = [100,100,100,100,100]. Is A > B? Mean(A) = 10k, Mean(B) = 100. Median(A) = 1, Median(B) = 100. Similarly Congress can, on average, beat the market while 2/3rds of them do not. I hope we see that the math even allows for us to approach 0% of members beating the market while an average value can. That’s a wildly different claim though. If you’re saying one person beat the market by a suspicious amount, you should say that. You should not say that a large group of people beat the market. The large group of people on average did not. Using the mean to describe a group dynamic when your mean is dominated by a single outlier is lying with statistics. I am a data scientist fwiw. > I am certain you know that "difficulty" is about intentional skill and not a reference to an act of god. You have removed any doubt that a conversation can happen in good faith. I don’t care about your semantic argument. It is very easy to beat the sp500 on a given year by random chance. I have done it for the last six years by a negligible amount. Do I have any particular insight? No. I’m 95% dumped in SPY and my 5% did a bit better. It is only beating it repeatedly and by a meaningful amount and net of any fees related to funds or investment managers that are hard. SPY is not a magic, difficult to beat goal. It’s simply a solid, average performing strat that takes zero effort. The reason most funds underperform it is because they do zany high risk stuff. If your portfolio is comprised of largely normal investments in common companies you’ll on average match the SPY with about half of your years beating it. > Clearly you do. But I actually don't. I don't wrestle with pigs. I really don’t. I don’t actually think politicians should be allowed to trade stocks at all for appearances alone. I just don’t like bad stats.

godelski 3y ago

> Congress beat the market absolutely doesn’t follow from 1/3 of them beating the market. It would be strange if that number were considerably lower. This is why I stressed considering what is marginalized out. I'll use an extreme example. Let A = [1,1,1,1,50000] and B = [100,100,100,100,100]. Is A > B? Mean(A) = 10k, Mean(B) = 100. Median(A) = 1, Median(B) = 100. Similarly Congress can, on average, beat the market while 2/3rds of them do not. I hope we see that the math even allows for us to approach 0% of members beating the market while an average value can. I am guessing the reason you disagree is likely one of two false assumptions: you've confused mean and median, or you've assumed performance is reasonably homogeneous (i.e. low variance). > No, that’s a stupid claim. It’s incredibly easy to beat the market in a single year by chance. I am certain you know that "difficulty" is about intentional skill and not a reference to an act of god. You have removed any doubt that a conversation can happen in good faith. > I don’t care man. Clearly you do. But I actually don't. I don't wrestle with pigs.

jncfhnb 3y ago
> This is why I stressed considering what is marginalized out. I'll use an extreme example. Let A = [1,1,1,1,50000] and B = [100,100,100,100,100]. Is A > B? Mean(A) = 10k, Mean(B) = 100. Median(A) = 1, Median(B) = 100. Similarly Congress can, on average, beat the market while 2/3rds of them do not. I hope we see that the math even allows for us to approach 0% of members beating the market while an average value can. That’s a wildly different claim though. If you’re saying one person beat the market by a suspicious amount, you should say that. You should not say that a large group of people beat the market. The large group of people on average did not. Using the mean to describe a group dynamic when your mean is dominated by a single outlier is lying with statistics. I am a data scientist fwiw. > I am certain you know that "difficulty" is about intentional skill and not a reference to an act of god. You have removed any doubt that a conversation can happen in good faith. I don’t care about your semantic argument. It is very easy to beat the sp500 on a given year by random chance. I have done it for the last six years by a negligible amount. Do I have any particular insight? No. I’m 95% dumped in SPY and my 5% did a bit better. It is only beating it repeatedly and by a meaningful amount and net of any fees related to funds or investment managers that are hard. SPY is not a magic, difficult to beat goal. It’s simply a solid, average performing strat that takes zero effort. The reason most funds underperform it is because they do zany high risk stuff. If your portfolio is comprised of largely normal investments in common companies you’ll on average match the SPY with about half of your years beating it. > Clearly you do. But I actually don't. I don't wrestle with pigs. I really don’t. I don’t actually think politicians should be allowed to trade stocks at all for appearances alone. I just don’t like bad stats.