S*******s
发帖数: 13043 | 1 就是说倍数大于1或小于0，随时间有确定的损耗，速率和波动平方成正比 |
o**o
发帖数: 3964 | 2 something might not be right
1. how about a=-1, constantly losing money?
2. a in (0,1)? |
B**********r
发帖数: 7517 | 3 Decay is proportional to
(1) square of underlying index
(2) square of 倍数
【在 S*******s 的大作中提到】
: 就是说倍数大于1或小于0，随时间有确定的损耗，速率和波动平方成正比
|
l****t
发帖数: 1379 | 4 大学教FINANCE的教授很多比不过TRADER挣的多, 就是因为老傻算, 以为自己很聪明.
其实没啥用.
If you want to make money, always buy low sell high.
就这么简单, 琢磨那么多干啥. 把简单的问题复杂化了.
【在 S*******s 的大作中提到】
: 就是说倍数大于1或小于0，随时间有确定的损耗，速率和波动平方成正比
|
l****y
发帖数: 8847 | 5 有才
那机构都可以关门了，玩不过散户
【在 l****t 的大作中提到】
: 大学教FINANCE的教授很多比不过TRADER挣的多, 就是因为老傻算, 以为自己很聪明.
: 其实没啥用.
: If you want to make money, always buy low sell high.
: 就这么简单, 琢磨那么多干啥. 把简单的问题复杂化了.
|
S*******s
发帖数: 13043 | 6 you dont know maths, do you?
【在 o**o 的大作中提到】
: something might not be right
: 1. how about a=-1, constantly losing money?
: 2. a in (0,1)?
|
S*******s
发帖数: 13043 | 7 I'm quant working for top tier traders in the world, who believe the
calculation worth more than they paid me.
【在 l****t 的大作中提到】
: 大学教FINANCE的教授很多比不过TRADER挣的多, 就是因为老傻算, 以为自己很聪明.
: 其实没啥用.
: If you want to make money, always buy low sell high.
: 就这么简单, 琢磨那么多干啥. 把简单的问题复杂化了.
|
S*******s
发帖数: 13043 | 8 obviously it is wrong because FAZ decays much faster than FAS
【在 B**********r 的大作中提到】
: Decay is proportional to
: (1) square of underlying index
: (2) square of 倍数
|
o**o
发帖数: 3964 | 9 It's not news that N x EFTs decay over time because the instruments they use
can't be gamma neutral, so lose money at constant re-balancing. Before
looking into the math, wanted to check some boundary conditions.
your last equation seems to suggest long and short the same stock (a=-1, b=0
?), the short leg constantly underperform the long?
【在 S*******s 的大作中提到】
: you dont know maths, do you?
|
l**i
发帖数: 8144 | 10 你这个东西的参数随时间变化么？
【在 S*******s 的大作中提到】
: 就是说倍数大于1或小于0，随时间有确定的损耗，速率和波动平方成正比
|
|
|
S*******s
发帖数: 13043 | 11 right.
use
=0
【在 o**o 的大作中提到】
: It's not news that N x EFTs decay over time because the instruments they use
: can't be gamma neutral, so lose money at constant re-balancing. Before
: looking into the math, wanted to check some boundary conditions.
: your last equation seems to suggest long and short the same stock (a=-1, b=0
: ?), the short leg constantly underperform the long?
|
S*******s
发帖数: 13043 | 12 你这个东西不会自己看？
【在 l**i 的大作中提到】
: 你这个东西的参数随时间变化么？
|
B**********r
发帖数: 7517 | 13 I think you are right. I just did some calculation. If the up/down days are
equal, and the underlying index returns to a previous value, FAZ decays
about 2X as fast as FAS. The rate of decay is still proportional to the
square of underlying index volatility.
It is a common misunderstanding if one believe that because the pair move
together, so they decay at the same rate.
I have another preference to short FAZ/TZA.
I also find 3X bull ETF decays as fast as 2X bear ETF (certainly assume they
are based on the same index). I don't know if your equation implies the
same.
【在 S*******s 的大作中提到】
: obviously it is wrong because FAZ decays much faster than FAS
|
B**********r
发帖数: 7517 | 14 Just more clarification based on the assumption of equal up/down days. For a
nX bull/bear ETF pair on an underlying index with volatility v. The rates
of decay are:
Bull ETF: ~ n(n-1)v^2
Bear ETF: ~ n(n+1)v^2
~ means proportional
are
they
【在 B**********r 的大作中提到】
: I think you are right. I just did some calculation. If the up/down days are
: equal, and the underlying index returns to a previous value, FAZ decays
: about 2X as fast as FAS. The rate of decay is still proportional to the
: square of underlying index volatility.
: It is a common misunderstanding if one believe that because the pair move
: together, so they decay at the same rate.
: I have another preference to short FAZ/TZA.
: I also find 3X bull ETF decays as fast as 2X bear ETF (certainly assume they
: are based on the same index). I don't know if your equation implies the
: same.
|
u********e
发帖数: 4950 | 15 好贴,顶
应该没有-1 到 1 之间的ETF 吧
【在 S*******s 的大作中提到】
: 就是说倍数大于1或小于0，随时间有确定的损耗，速率和波动平方成正比
|
B**********r
发帖数: 7517 | 16 For a 1x ETF, n=1, there is no decay if it is a bull ETF, but it has decay if it
is a bear ETF.
【在 u********e 的大作中提到】
: 好贴,顶
: 应该没有-1 到 1 之间的ETF 吧
|
u********e
发帖数: 4950 | 17 对
* 1x should have NO decay since it is just pure long stocks, and it should
also has dividend income
* -1x: should have NO decay since it is just pure short stocks, but it does
need to pay interest and dividends ( or, could be different kind of decay
compared to what LZ said)
* -1 to 1: could be achieved by a comination of cash position and 1x or -
1x. So there should be no decay also (except the interest and dividends )
it
【在 B**********r 的大作中提到】
: For a 1x ETF, n=1, there is no decay if it is a bull ETF, but it has decay if it
: is a bear ETF.
|
B**********r
发帖数: 7517 | 18 An ETF may do daily rebalance. Then a bear ETF is no longer a pure short
play.
should
does
decay
【在 u********e 的大作中提到】
: 对
: * 1x should have NO decay since it is just pure long stocks, and it should
: also has dividend income
: * -1x: should have NO decay since it is just pure short stocks, but it does
: need to pay interest and dividends ( or, could be different kind of decay
: compared to what LZ said)
: * -1 to 1: could be achieved by a comination of cash position and 1x or -
: 1x. So there should be no decay also (except the interest and dividends )
:
: it
|
u********e
发帖数: 4950 | 19 Yeah, the "daily rebalance" is a evil here.
【在 B**********r 的大作中提到】
: An ETF may do daily rebalance. Then a bear ETF is no longer a pure short
: play.
:
: should
: does
: decay
|
B**********r
发帖数: 7517 | 20 Let us say, the index is up x one day, and down x/(1+x) the following day.
The index returns to (1+x)(1-1/(1+x)) = 1
The 1X bear index is down x first, and then up x(1+x) the second day. The
ETF returns to (1-x)(1+x/(1+x)) = 1-2x^2/(1+x), a decay.
should
does
decay
【在 u********e 的大作中提到】
: 对
: * 1x should have NO decay since it is just pure long stocks, and it should
: also has dividend income
: * -1x: should have NO decay since it is just pure short stocks, but it does
: need to pay interest and dividends ( or, could be different kind of decay
: compared to what LZ said)
: * -1 to 1: could be achieved by a comination of cash position and 1x or -
: 1x. So there should be no decay also (except the interest and dividends )
:
: it
|
|
|
o**o
发帖数: 3964 | |
B**********r
发帖数: 7517 | 22 Thanks for the good link. It confirms what I calculated.
【在 o**o 的大作中提到】
: http://en.wikipedia.org/wiki/Inverse_exchange-traded_fund
: also its reference at http://olympiainv.com/Memos/ETFs.pdf
|
l****g
发帖数: 5080 | 23 这个没什么错误，所以ETF不能hold太长时间，几个星期就已经长了，例如FAZ之类。金
银的ETF似乎可以hold个一两年。 |
B**********r
发帖数: 7517 | 24 You must be kidding. Look at AGQ/ZSL's chart YTD together, and you know all those
ETFs are the same scams.
【在 l****g 的大作中提到】
: 这个没什么错误，所以ETF不能hold太长时间，几个星期就已经长了，例如FAZ之类。金
: 银的ETF似乎可以hold个一两年。
|
B**********r
发帖数: 7517 | 25 to top
a
【在 B**********r 的大作中提到】
: Just more clarification based on the assumption of equal up/down days. For a
: nX bull/bear ETF pair on an underlying index with volatility v. The rates
: of decay are:
: Bull ETF: ~ n(n-1)v^2
: Bear ETF: ~ n(n+1)v^2
: ~ means proportional
:
: are
: they
|
