founder @bprintsfi @touch_market; prev. some "head of" roles; I like math. (triple olympiad medalist); @paradigm fellow
Abstract
Joined May 2022
- Tweets3.2K
- Following1.7K
- Followers1.8K
- Likes7.2K
A mainstream (MiCA-regulated) CEX keeps changing the amount I *paid* on a trade hours after it completed.
I caught half a dozen different numbers on screenshots.
Contacted support. Tried 5 times (that's how many times I had to explain) before they actually created a ticket.
I think it's people's role to find bulishness on things
my role is to point out what to be bearish on
and it doesn't seem to be very appreciated on X
should I stop pointing things out?
My take on how papertrade will end up:
Whoever thinks zero-slippage trading is possible is simply misunderstanding what markets are about.
There's always expected value to be donated to whoever knows a little more about how the price will move.
People do put some value in paper, though, so this is imo the most likely way it will play out:
1. The house funds will get depleted (most likely relatively slowly, for a few days or weeks) until $2m
2. At $2m people will try to donate to the house to farm paper. As long as paper is relatively valued, this will hold. But it's not a never-ending protection! Paper holders will get diluted until a point where this fails further an further.
3. Not having a public venue to trade paper is a bliss right now. If there was one, all the farmers would try to sell and the market would discover that it's nearly valueless aside the ability to farm itself. That would give a very good argument not to lose to the house to farm paper – in the end, it's better to just buy paper if you think the value would be higher.
4. At some point, there will be no one willing to "buy" newly issued paper for dollars, and that's when the manipulation/good trader leak will overcome and make the product insolvent.
This is already visible. The cheapest you can get paper is 100 paper/$. Usually it's ~3x cheaper. How, with ~6b paper outstanding at the time of writing, that's AT LEAST $60m in losses that generated the paper. The product captured $26.3m to buybacks and $3.5m to house funds. Where did the remaining funds go?
They were extracted already. $30m+ of funds already leaked from the system and we should expect more to come. At a point the market will realize this leak is not worth having exposure to and paper will become nearly worthless.
The house balance will go negative, sometimes even payout itself out, until it doesn't.
Just think what happens if the system stops minting paper (i.e. it becomes ± worthless).
Don't get me wrong, paper can make some money on actual clients coming to gamble/trade, but imo it's unrealistic a few days after launch. Biggest part of current usage is paper farming. The value leak can be so fast that I think it's extremely risky (imo improbable) to assume revenue will just overtake it.
It was a long-known issue with papertrade and founders have no intention on fixing this even after seeing it in action. I'd have no hopes investing in paper based on a thesis that something will change.
This bears some to the LUNA<>UST<>USDC situation, but with PAPER<>deposited USDC<> USDC.
Stay safe out there and away from solvency gambling.
Just watched. Highly recommend if you're a STEM person!
Some reflections on AI in math. I found it helpful to simply list out the various parts of what a mathematician does, why math is valuable, and where the funding comes from. From there, what happens if AI gets more capable?
youtu.be/LKrwhjOI8Nc
Papertrade daily update #1:
~600m more paper outstanding
~$0.5m leaked from the house
~$4.9m in staking rewards
implied paper price: $0.0073
if it's supposed to be greater, papertrade leaks value
My take on how papertrade will end up:
Whoever thinks zero-slippage trading is possible is simply misunderstanding what markets are about.
There's always expected value to be donated to whoever knows a little more about how the price will move.
People do put some value in paper, though, so this is imo the most likely way it will play out:
1. The house funds will get depleted (most likely relatively slowly, for a few days or weeks) until $2m
2. At $2m people will try to donate to the house to farm paper. As long as paper is relatively valued, this will hold. But it's not a never-ending protection! Paper holders will get diluted until a point where this fails further an further.
3. Not having a public venue to trade paper is a bliss right now. If there was one, all the farmers would try to sell and the market would discover that it's nearly valueless aside the ability to farm itself. That would give a very good argument not to lose to the house to farm paper – in the end, it's better to just buy paper if you think the value would be higher.
4. At some point, there will be no one willing to "buy" newly issued paper for dollars, and that's when the manipulation/good trader leak will overcome and make the product insolvent.
This is already visible. The cheapest you can get paper is 100 paper/$. Usually it's ~3x cheaper. How, with ~6b paper outstanding at the time of writing, that's AT LEAST $60m in losses that generated the paper. The product captured $26.3m to buybacks and $3.5m to house funds. Where did the remaining funds go?
They were extracted already. $30m+ of funds already leaked from the system and we should expect more to come. At a point the market will realize this leak is not worth having exposure to and paper will become nearly worthless.
The house balance will go negative, sometimes even payout itself out, until it doesn't.
Just think what happens if the system stops minting paper (i.e. it becomes ± worthless).
Don't get me wrong, paper can make some money on actual clients coming to gamble/trade, but imo it's unrealistic a few days after launch. Biggest part of current usage is paper farming. The value leak can be so fast that I think it's extremely risky (imo improbable) to assume revenue will just overtake it.
It was a long-known issue with papertrade and founders have no intention on fixing this even after seeing it in action. I'd have no hopes investing in paper based on a thesis that something will change.
This bears some to the LUNA<>UST<>USDC situation, but with PAPER<>deposited USDC<> USDC.
Stay safe out there and away from solvency gambling.
My take on how papertrade will end up:
Whoever thinks zero-slippage trading is possible is simply misunderstanding what markets are about.
There's always expected value to be donated to whoever knows a little more about how the price will move.
People do put some value in paper, though, so this is imo the most likely way it will play out:
1. The house funds will get depleted (most likely relatively slowly, for a few days or weeks) until $2m
2. At $2m people will try to donate to the house to farm paper. As long as paper is relatively valued, this will hold. But it's not a never-ending protection! Paper holders will get diluted until a point where this fails further an further.
3. Not having a public venue to trade paper is a bliss right now. If there was one, all the farmers would try to sell and the market would discover that it's nearly valueless aside the ability to farm itself. That would give a very good argument not to lose to the house to farm paper – in the end, it's better to just buy paper if you think the value would be higher.
4. At some point, there will be no one willing to "buy" newly issued paper for dollars, and that's when the manipulation/good trader leak will overcome and make the product insolvent.
This is already visible. The cheapest you can get paper is 100 paper/$. Usually it's ~3x cheaper. How, with ~6b paper outstanding at the time of writing, that's AT LEAST $60m in losses that generated the paper. The product captured $26.3m to buybacks and $3.5m to house funds. Where did the remaining funds go?
They were extracted already. $30m+ of funds already leaked from the system and we should expect more to come. At a point the market will realize this leak is not worth having exposure to and paper will become nearly worthless.
The house balance will go negative, sometimes even payout itself out, until it doesn't.
Just think what happens if the system stops minting paper (i.e. it becomes ± worthless).
Don't get me wrong, paper can make some money on actual clients coming to gamble/trade, but imo it's unrealistic a few days after launch. Biggest part of current usage is paper farming. The value leak can be so fast that I think it's extremely risky (imo improbable) to assume revenue will just overtake it.
It was a long-known issue with papertrade and founders have no intention on fixing this even after seeing it in action. I'd have no hopes investing in paper based on a thesis that something will change.
This bears some to the LUNA<>UST<>USDC situation, but with PAPER<>deposited USDC<> USDC.
Stay safe out there and away from solvency gambling.
signing ethereum-style transactions for hft is fascinating
naive, hyperoptimized implementations do it in like 10µs
but there's magic to be done
you can prepare for signing a message without knowing what you'll need to sign
i'll repeat
*without knowing what you'll sign*
so, you can also prepare yourself to sign many things
doing it for many signatures actually makes it way quicker, even 2x --> 5µs
but that's nothing compared to the next fact
after getting to know what you's signing, you just have to do <1% of the remaining work
yes
>99% of the whole signature computation can be done without knowing what you're signing
so after preparing everything, you can "burst" sign MANY things and it costs you nearly nothing
<0.02µs
not a typo
now, preparing signatures can be easily delegated to another thread
or multiple threads
you're never waiting for other threads
they're only doing their work and filling up buffers of signature precomputes for you
you just read them without synchronization
and magically, you can sign with lower latency and more throughput than you could ever want
pretty beautiful things happening in this world
you may ask
"if we do all this computation without knowing what we're signing, this means it can be used for any message"
"why don't we just compute it once?"
that's a great thought
but unfortunately ecdsa becomes unsafe then
you can only use a single "prepared" payload once, otherwise
boom
you're hacked
stay safe out there
