Fireball (OP)
|
|
November 30, 2012, 10:23:16 PM |
|
Wow, that was a fast discovery. I added it to the list and put first orders to test it, like, 5 or 10 minutes ago :-)
|
|
|
|
ThePok
|
|
November 30, 2012, 10:41:37 PM |
|
Yeah, maybe becaouse iam your biggest fan i did place the first real order! i want 100 of that lovly shares
|
|
|
|
guruvan
|
|
December 01, 2012, 08:07:15 PM |
|
Sell Price Buy 20 12.84 56 12.75 10 12.74 25 12.73 L 1 12.49 12.48 5 12.45 6 12.42 42 12.41 18 12.401 22 Range 11.2365 - 13.7335 It is really quite inexcusable that the 1@12.49 placed moments before clearing affects the clearing price as much as orders for considerable sums that have been up for long periods of time. Bullshit.
|
|
|
|
Fireball (OP)
|
|
December 01, 2012, 08:42:26 PM Last edit: December 01, 2012, 08:52:29 PM by Fireball |
|
Sell Price Buy 25 12.73 L 1 12.49 12.48 5
The spread is higher than usual so it's easy for users to insert orders inside of it. If you don't want this to happen, just add a buy order at a higher price :-) Anyway, after we're done with existing bunch of updates, I plan to move on to popularizing automated arbitrage (a new topic will be started for its discussion, I think) and share the already developed code for node.js and C# trading bots supporting ICBIT+MtGox. This should bring profits to users running bots thanks to arbitrage, and to futures market players because of higher liquidity available at a given moment and faster price changes (ultimately opening up scalping possibilities).
|
|
|
|
picobit
|
|
December 02, 2012, 12:53:57 PM |
|
It is really quite inexcusable that the 1@12.49 placed moments before clearing affects the clearing price as much as orders for considerable sums that have been up for long periods of time. If something like that gives you a loss today, it will give you a corresponding winning tomorrow at the next clearing - or when you close your positions. So in the long run, it has no effect unless your positions are forced close, which of course could happen.
|
|
|
|
Stephen Gornick
Legendary
Offline
Activity: 2506
Merit: 1010
|
|
December 02, 2012, 06:13:17 PM Last edit: December 02, 2012, 08:54:31 PM by Stephen Gornick |
|
So in the long run, it has no effect unless your positions are forced close,
Or you put in a lowball bid and just wait for "8 O'clock Charlie" to strike. Incidentally, this works both ways. We've seen buying right before the bell that pushes the clearing price up above market as well. Liquidity is the solution. And these type of profit opportunities are just what brings more liquidity.
|
|
|
|
Stephen Gornick
Legendary
Offline
Activity: 2506
Merit: 1010
|
|
December 02, 2012, 09:22:19 PM |
|
"Flash crash" is not possible, because short spikes won't produce a mass position liquidation in our system, especially if it goes *against* spot market. However, don't forget: if you're in "margin call zone", liquidation of your open positions may happen at every moment. Either fill up your account as soon as you can, or face consequences.
Those read like two conflicting statements. Am I misunderstanding? To me it says. A huge price swing won't cause liquidation. But be careful because a huge price swing could cause liquidation of your positions. The reason I am wondering is perhaps I would like to manage my balance to not have significantly more than is necessary to meet the margin maintenance requirements. If I simply needed to have sufficient maintenance at the time of clearing, then that would be easy -- as I could make sure to add more funds or do some trades before clearing if it looked like I might be getting close to having insufficient margin remaining. But if I have to worry that at any point in time throughout the day if a price swing goes against me my positions could get liquidated then I'm going to be more focused on making sure my margin maintenance level never ends up being low enough to where I'm exposed to that risk. Is there more info on how this (liquidations) are handled?
|
|
|
|
Fireball (OP)
|
|
December 02, 2012, 09:59:01 PM |
|
"Flash crash" is not possible, because short spikes won't produce a mass position liquidation in our system, especially if it goes *against* spot market. However, don't forget: if you're in "margin call zone", liquidation of your open positions may happen at every moment. Either fill up your account as soon as you can, or face consequences.
Those read like two conflicting statements. Am I misunderstanding? To me it says. A huge price swing won't cause liquidation. But be careful because a huge price swing could cause liquidation of your positions. Nope. The first line describes why there is a price boundary for every trading session. If there would be none, then someone with 700 BTC (as of now) could buy all sell-side of the orderbook and become the one who defines current market price. Nothing would stop from putting orders like $1000 per 1 BTC. The second line states: Even though the price fluctuations are limited within one trading session, it means you still have to check and ensure that you have enough money if the price goes against you. Also, it states that you are informed that if there is not enough money, then forced sell (forced buy) may happen at any moment. It does not mean it's going to happen. All margin calls are being human-reviewed now, and only extreme cases are being force closed, often after unsuccesfully trying to reach the person by email beforehand.
|
|
|
|
guruvan
|
|
December 05, 2012, 01:16:37 PM |
|
It is really quite inexcusable that the 1@12.49 placed moments before clearing affects the clearing price as much as orders for considerable sums that have been up for long periods of time. If something like that gives you a loss today, it will give you a corresponding winning tomorrow at the next clearing - or when you close your positions. So in the long run, it has no effect unless your positions are forced close, which of course could happen. It affects available cash today once VM is assessed at clearing time, so it really does actually matter that it's not fucked every other day. (and, yeah, I'm talking about movement in either direction, caused by insignificant orders placed seconds before clearing...placed too late for anyone else to hit the offending bid/ask) I maintain that it's bullshit that an order for one contract, placed seconds before clearing time, moves the clearing price as much as 10 contracts sitting there for hours. The fact is, the movement it creates is artificial, and it affects the trading that occurs in the next session because of the real or psychological effect of the Execution price in the traders' portfolios.. It is not, as picobit says, the case that "a loss today will be a corresponding winning tomorrow" Not even close. I believe Fireball misunderstood the complaint. Placing a higher bid was not possible in the seconds available. The issue was the effect despite the proximity to clearing time. I have no idea the relevance of the arb-bot to this issue.
|
|
|
|
Fireball (OP)
|
|
December 05, 2012, 03:03:48 PM |
|
Placing a higher bid was not possible in the seconds available.
Why can't you place a higher bid beforehand? There is no need to submit orders in the last second, all limit buy and sell orders are kept in the system until canceled.
|
|
|
|
Ichthyo
|
|
December 06, 2012, 02:37:11 AM |
|
The ICBIT pricing engine has clearly some kind of shortcoming, which can be used to game the settlement price in any futures market with low liquidity. As I've proposed some time ago already, I think it would be an interesting experiment how those markets would develop if the engine used the weighted average of the last trading session for the daily settlement. Would those markets dry out, because being completely boring? or would such a change rather increase the inflow, since they are less risky then? Bitcoin is a nice field for experiments.
But I understand that Fireball tends to stay away from such experiments and prefers to keep the platform rather close to the behaviour of the futures known from the conventional finance industry.
Anyway, whenever our 8'o clock Charley succeeds with gaming the rate away, this creates nice arb possibilities in the next trading session(s). And, obviously, if there was a whole bunch of trading bots just waiting to catch such possibilities, the problem would be mostly gone.
Bottom line: while I'm not completely happy with the behaviour of the engine, at least it is predictable, and in the end, the possibility of such (limited) manipulation games is for me just part of the risk equation.
|
|
|
|
guruvan
|
|
December 07, 2012, 07:02:27 PM |
|
True.
On further consideration, I suppose I wouldn't really notice this action if it was (considerably) more liquid, and there were more orders placed right at clearing time (that being the primary difference between icbit and current fiat-based futures exchanges)
|
|
|
|
segabtc
|
|
December 09, 2012, 05:03:25 PM |
|
Noticed this a day or two ago, when I cancel an order in the "Your Orders" window, it cancels the order but also I lose all other orders info on that window til I "reload" the page. Can we get a "reload" button like the "Balances" window has? Instead of reloading entire page, just something that will refresh or reload that one window? Will save alot of bandwidth and server load. Again, I know it is the holiday season and this might not get done for a few weeks. Just suggestions for a better user experience for all. If not tell me to STFU! heheheh
|
|
|
|
killerstorm
Legendary
Offline
Activity: 1022
Merit: 1033
|
|
December 09, 2012, 05:14:02 PM |
|
it cancels the order but also I lose all other orders info on that window til I "reload" the page. Can we get a "reload" button like the "Balances" window has?
It doesn't need a reload button, bug which wipes your orders should be fixed instead.
|
|
|
|
segabtc
|
|
December 09, 2012, 05:29:30 PM |
|
ok that would work also!
|
|
|
|
Fireball (OP)
|
|
December 09, 2012, 08:35:09 PM |
|
ok that would work also!
I'm gonna check, thanks for reporting. In fact, I introduced a change in the orders table updating algorithm a few days ago, so probably it's related to that.
|
|
|
|
segabtc
|
|
December 09, 2012, 08:52:29 PM |
|
damn skippy, love it when a developer knows his sh|t!
|
|
|
|
guruvan
|
|
December 10, 2012, 01:24:06 AM |
|
FYI, Fireball: - memory leak issue seems to still be pretty serious - seems since I mentioned memory the first time - now whenever I place or cancel any order, all my pending orders disappear from their frame until reload the page. Thanks for looking into it!! EDIT: yea.. see more above with the same issue.
|
|
|
|
Stephen Gornick
Legendary
Offline
Activity: 2506
Merit: 1010
|
|
December 11, 2012, 02:27:38 AM Last edit: December 11, 2012, 02:40:48 AM by Stephen Gornick |
|
2. Because of that, trading participants will try to "arbitrage" this fact and will move the price quite close to that value before trading stops.
And that's exactly what's happening. So the BUH3 (March 15, 2013) BTC/USD futures are in la-la land at $16 and change: - https://icbit.se/BUH3While BUZ2 (December 15, 2012) is now within 1% of the Mt. Gox last / spot market price with just a few more trading days remaining.
|
|
|
|
smickles
|
|
December 11, 2012, 04:01:17 AM |
|
2. Because of that, trading participants will try to "arbitrage" this fact and will move the price quite close to that value before trading stops.
And that's exactly what's happening. So the BUH3 (March 15, 2013) BTC/USD futures are in la-la land at $16 and change: - https://icbit.se/BUH3While BUZ2 (December 15, 2012) is now within 1% of the Mt. Gox last / spot market price with just a few more trading days remaining. 16 is lalaland? I was trying to buy it up to 50 b/c I fully expect it to be there in march jk
|
|
|
|
|