813900+ entries in 0.605s
mike_c: which is still only a piece of it. because if someone drops
that 1k bet down on my side my odds still
tank
ozbot: Blancmange curve - Wikipedia,
the free encyclopedia
mike_c: mircea_popescu: yeah, i was
thinking about
that. but its not like its not a cloudy soup right now.
jurov: just
throw weierstrass function in
jurov: 1+cos(x) in
the 0,pi interval? but yes, needs some modelling what would be
the effect
mircea_popescu: mike_c one major problem of
the not-time approach is, what if right before your 10btc bet makes it, someone puts in a 1k btc bet.
Vexual: u in it for
the money or in it for
the love?
mircea_popescu: "any woman would be better
than
the girl next door. except, you know, a really annoying pornstar. or a quadriplegic. or a butch lesbian."
ThickAsThieves: i mean
that in
the starting from linear sense, not saying an extreme
taper would be better
Vexual: add some function
to
the diff estimate
to make a sexy curve
Vexual: perhaps
the weather curve is plotted from data
Vexual: a curve from
the difficulty estimate?
ThickAsThieves: what if it's a fixed amount of weight
that can be given,
then
the 'price' of weight is determined by
the amount of coins bet prior.
mircea_popescu: however, it may be
the best approximation, depending how
the problem is actually defined
mircea_popescu: i mean
to be clear, im not proposing linear
timeweight is
the absolutely correct solution. it's obviously wrong, life isn't linear.
ThickAsThieves: if it's not on
the agenda for
the gov by a certain date
mircea_popescu: Vexual it's where people seem
to
think
they do a lot better at predicting outcomes as
time goes by
Vexual: you're only
talking about difficulty bets?
mike_c: yeah, i'll
think it
through.
mike_c: thinking it
through. i believe it is different because
the guy who bets
the first 10btc gets more
than
the guy who bets
the last 10btc
mircea_popescu: mike_c how is your idea different from just splitting
the pool according
to a polynomial (bet sum) ?
mike_c: only marginally. but if you auction off weight, it changes
the dynamic a lot
mircea_popescu: ThickAsThieves basically he proposes
to not use a
time function, but a bet function
mike_c: if you "buy" weight, it encourages you
to 1) wait for more info, but 2) get in before someone else gets all
the good weight
mircea_popescu: i can follow mike_c proposal but not yours, and moreover
they seem quite distinct
ThickAsThieves: then each bet get 10% of
that weight, but new bet's weight is offset
mircea_popescu: jurov a nonlinear curve would, of course, have
to somehow be communicated
to
the bettors, which may
take 16 months of scam accusations in
the btctalk.
mike_c: like, you bet 50btc,
the weight drops 50k
mike_c: it puts something of a cap on
the bets, but whatever.
mike_c: i'm warming up
to
the 'spend weight' idea
jurov: and i strongly dount
that curve is linear
jurov: hm,
there should exist some curve how
the prediction accuracy increases over
time for natural phenomena
mircea_popescu: what has materially changed in
that bet since a month ago ?
ThickAsThieves: of course
then you run into
the other issue of letting people make losing winning bers
mircea_popescu: not only what mike_c said, but also...
then how do you specify
the curve.
mike_c: not sure
that's any different
than just closing early
jurov: yes, updating
the curve could help
jurov: no, all
the bets i can bet, but have weight like 50k or 30k
mircea_popescu: so
then what dio you do, have all
these bets you can't bet on ?
mircea_popescu: "Even with weights, allowing a diff bet
to be open
to bets
till so close
to
the resolution date, enormously reduces
the incentive
to bet until it is an obvious call and
thus limits
the size of
the bets
to a few eager early beters plus roughly double
the late skimmers. No money in between an
thus smaller overall rollers."
mircea_popescu: ok so basically
they mostly delivered an avalon clone on a 9 month schedule, and announced a 28nm process on a 3 month schedule late sept.
ozbot: Official Avalon mining rig clones
thread from bitmine.ch (NOW SHIPPING)
mircea_popescu: so basically
this is a 25% actual rigs (from cointerra) and 75% vaporware.
Duffer1: no i'm not sure if
they've actually shipped
though
Duffer1: a bunch of spam mail
to my inbox but
that's about it
the20year2: It's not like you can just find a datacenter and say "hey let me put my stuff in
there"
the20year2: My brother's looking at local warehouses/factories
to deploy in, and
the cost for setup and
time involved is immense
the20year2: I wonder what kind of facility
they're in
mircea_popescu: 240
th
they say
they're deploying, selling
the lot for 50k btc.
the20year2: *Hopefully*
the end result will be smarter investors
ThickAsThieves: last i checked,
taking candy from babies wasn't good for
the rep
Duffer1: it's not intended
to benefit
the investor so i'd call
that a scam, maybe not "theft"
Vexual: courts oft call it
theft
Duffer1: fleecing
the ignorant is little better
than
theft imo
the20year2: is it
theft if
they willingly hand it over?
mircea_popescu: stealing 2.8mn from idiots does not make
the stolen funds investments.
Duffer1: what's your project
the20?
the20year2: makes me wish i'd of asked for 5k bitcoins instead of
the 700
the20year2: Who cares if
they get called a scam?
They've already pocketed 2.8m in investments