excluding the first minute.
1,000
1.00I posted about this market before and noted that it seemed like the ideal strategy (until closer to expiry) was just to place a limit order on yes at 81%. I myself did that and it worked out perfectly. As expected, once @JonathanRay cancelled his limit orders, the market did not stay above 81% for long so it resolved yes. The only thing I regret was I didn't have much free mana in my account when this market was resolved overnight, so only a relatively small amount actually executed for me at 81%. But it was a free 24% profit overnight when I wasn't even paying attention.
The one thing that's a little funny about it is it seems like @khang2009 traded irrationally. Around midnight, he bought 8,700 "no", pushing the market to 80%, which guaranteed that his "no" position would immediately become worthless. So he lost 7.6k on this market, the most of anyone. The only way that is economically rational is if he's working with @JonathanRay but that doesn't look to be the case. Since he incurred the losses of clearing out the limit orders at 81%, @JonathanRay didn't have to so @khang2009 effectively transferred mana to everyone who had a net long and the folks who had limit orders at 81%.
In any case, fun times were had by all!
/ bought Ṁ8,700 NO from 81% to 80% 9h
so resolves yes.
@JonathanRay
Meanwhile Free 3%
So excluding Jonathan Ray, amount he needs to sell has doubled and his position looks less profitable as a result of this increase. Will this trend scare Jonathan Ray into doing it sooner rather than waiting?
So despite that doubling, it still seems likely to resolve yes unless a lot more buy orders at 81% added. (They should be added by all people holding no positions that haven't already covered their position by doing so). In particular Niel S ( @NielS ) should still add a lot more buy orders at 81% to protect his position.
You are right, i put up some more limit orders. But how i see it, say you buy no shares at a good limit price. Then you put a yes limit order at the same time at 81 cent. The market can only go below 81 when all orders are filled. So it's like a hedge at that point. You own no and it will resolve no. Or market goes to yes and orders get also filled flipping the position around
@NielS Yes you have put some more up. But I a still confused. You now have a no position of 32231 at a cost 5826 mana (average cost 0.18076 )
You currently have orders at 81% for 537+500+500+500+250+3000+300+500 = 6037 mana which would cover a position of 7453
So if Jonathan Ray sold it down to one mana at 80% then he would buy also all the no orders but leaving those aside for the moment you would have spent 5826 and 6037 and only get back 7453 for a loss of 4410 mana. This seems to me to be likely.
Whereas if you increased your yes orders at 81% from 6037 mana to 26107 you would be fully protected. Either it resolves no and your no position makes great return or it goes down to 81% and you still make 1% on your 32231 position though likely losses on any further no orders you have outstanding at the time Jonathan Ray makes his move..
So as far as I can see your yes orders at 81% should be at least 26107 mana rather than a mere 6037
I think optimal strategy is this: Bet "yes" at 81% for now. If @JonathanRay keeps his limit orders up, then closer to expiry, you might switch to no. This market is basically just about game theory with @JonathanRay .
Here's why:
As of this writing, Jonathan is holding ~40k manna bet on 'Yes" and has limit orders for more than 1M manna on "yes" at 81% (his actual balance is only 595k, however, so beyond that point those limit orders are "fake" and would not actually execute only his balance is exhausted). It seems to me that betting "no" against Jonathan at this early stage is a bad bet, because his incentive is just to continue accumulating a "yes" position as long as he can -- then sometime before expiry, cancel all his limit orders at 81%, execute against any other limit orders others have placed on "yes" and push the market briefly to 80%, which means his net "yes" position immediately pays off. He can do this anytime from now until November, so for now, your best play is to place a "yes" limit order at 81% -- which will force him to essentially pay you off if he does as I expect and later pushes the market to 80%.
However, if he keeps his "yes" limit orders up close to expiry, you could earn even greater profits at his expense (though it works best if a consortium works together): If we're close to the end and the market has stayed above 81%, then you can execute against his "yes" orders for about 20% of your bankroll. (Be careful to leave some liquidity at 81% so the price stays above that. Keep an eye on his balance so you stop executing before it falls to zero.) Then immediately place "yes" orders at 81% with your remaining bankroll (roughly four times the size of whatever "no" order you placed at 81%). If everyone follows that strategy (or even one or two players with a large bankroll does it), then in order to push the market below 81%, he'll have to take the opposite side on your "yes" orders (by betting "no" at 81%) -- and his losses from doing that will more or less exactly offset his profit on the "no" order you placed at 81%. Plus, assuming you do this with a group close to the end of trading, you could nearly exhaust his balance at 81% (being careful not to fully exhaust it), which could limit the extent to which he can place "no" orders against you.
If Jonathan ultimately has enough liquidity to push the market below 81% before trading closes, you will lose a small amount due to transaction costs (assuming you calculated it correctly and your "yes" bet was sized to exactly offset your "no" bet, both of which would execute at 81%). However, if enough players adopt this strategy at the end (or at least one very large on does) and Jonathan doesn't have sufficient liquidity to fully offset his "yes" position in time, you'll quadruple your money at his expense on any portion of the "yes" bets that he's stuck with and can't offset. A major risk is that, at some point, he'll cancel his "yes" orders and you'll have to defend the peg with your own limit orders -- but a group of people doing this close to expiry could limit that risk, unless you think Jonathan is similarly colluding with his own group, which might come in with much more liquidity than he is displaying. :-)
Given that he is very vulnerable close to expiry, my guess is he'll just drop the peg sometime before then so keeping a limit order yes at 81% is a good bet. Just make sure your orders are small relative to his, because if you end up taking on a big "yes" exposure, then you'l be in the same position he is in.
Also, if you want, you can just bet your full balance (whatever Manifold will allow you to place) on a limit order at 80%, might as well. Unlikely that you'll ever see an execution of any material amount, but if someone makes a fat finger error while manipulating the price down to resolve "yes," it's free money.
I am surprised how little there is on yes at 81% other than Jonathon. He could remove his orders and sell just enough to get it to 80%. Only around 8k currently so leaves him with a large yes balance and profit.
Surely Neil S should put in orders to buy out his 29772 position at 81% to protect himself against this? (He only has orders for 2500 at 81% at present.)
Makes me wonder if Neil S and Jonathon Ray are working together?
If I was Jonathon and not working with Neil I would be worried that Neil would realise and put in the orders, so would likely push it down to 80% now. But perhaps Jonathon thinks leaving it longer will allow him to build a larger position?
Currently, he probably would not profit by dropping the market to 80% (assuming he's no colluding with others). There are Ṁ13,377 in limit orders at 81% from other traders; his maximum payout on his position is listed as Ṁ67,933, but he has already spent Ṁ55,723. To shift the market to 80%, he would have to spend an additional Ṁ13,377 (+1) to execute against all the "no" orders, bringing his expenditure to Ṁ69,101, which results in a loss of Ṁ1,168. As the market organizer, he receives half of the fees from executions, which might bring him to break-even or a slight profit, but not a significant one (additionally, as he places large market orders, Manifold may introduce some "liquidity," meaning orders opposite to his, requiring him to spend more than estimated to clear all the "no" orders and push the market to 80%).
One if too many people place limit orders at 81% is that you have to be ready to cancel if @JonathanRay cancels his. My analysis of the incentive structure above assumes he intends to maintain his 81% limit orders for some long period. However, if too many limit orders are placed at 81%, he might opt to cancel his to avoid increasing his position and hope that others are less inclined to defend the peg, potentially allowing the market to drop below 81% due to others' actions, enabling him to profit from his existing "yes" position without incurring losses.
To put it another way, there's a bit of a collective action problem among traders who actually acquire a position in "yes." At some point people who are heavily net "yes" will need to incur losses by executing against any "yes" limit orders at 81% to push the market to 80% and profit on their net "yes" position. But everyone holding "yes" would prefer for someone else to do that. Also, "yes" is not that profitable since it's priced at a minimum of 81% to start, so even if you win on "yes", you earn a return of at most 23% (= 1.00 / 0.81) on what you invested.
I appreciate @JonathanRay for creating this market. It presents a fascinating game theory and coordination challenge.
I see orders for 8877 slightly higher than the 8k I mentioned earlier. Grossing this up comes to 10955. So not sure where your 13377 is coming from unless more have been added then deleted.
The obvious strategy to me is to sell at above ~83% then put in buy order at 81% to cover your position.
His average purchase price is 82% So lose 1% on 11000 to gain 18% on the remaining 58k sounds like a nice 10k profit to me.
But if Neil S carries out my strategy above of covering no position by adding an extra 26k of yes orders at 81% then it becomes lose 1% on 37k to gain 18% on 32k. Still a good profit but only 5k rather than 10k.
Yeah, the orders are changing. Currently I see Ṁ8,627 in "yes" orders (other than @JonathanRay ) at 81%.
Your profit calculation might be off somewhat. You said, "lose 1% on 11000 to gain 18% on the remaining 58k." He would lose 100% (not 1%) on the Ṁ8,627 to gain 18% on the remaining 58k. The numbers Manifold has up says Jonathan Ray spent Ṁ55,723 for a potential max payout of Ṁ67,933, so potential profit of Ṁ12,210 if someone else pushes the market to 80%. But since he has to eat losses of Ṁ8,627 to do it himself, his profit would be only Ṁ3,582 if he does it now (not much in the context of his balance). If Neil S does what you say, he'd lose > Ṁ22k to force market to 80%.
Selling "no" at 83% and buying "yes" at 81% doesn't seem profitable because you pay a fee of more than 5% when you sell "nos" at that level and you also are unlikely to obtain execution of both sides of the trade (certainly not quickly or close to simultaneously). You're depending on someone else acting irrationally. Even if you manage to obtain execution on both sides through limit orders (so you don't pay the transaction fee), you're only gaining 2.4% and that could require tying up mana for months if you don't get execution on both legs right away, so profit is below the risk-free interest rate. There's also some risk that you only see execution of your "yes" orders at 81%, and then the market stays at 82% until close, in which case you lose 100% of your investment, so it's not risk free. Obviously this isn't real money, but if it were, you'd be strictly better off just holding Treasuries.
5% in fees could well make 83% to 81% loss making. But there is very little risk of loss if you don't place the yes order til after you get the no order filled.
>"He would lose 100% (not 1%) on the Ṁ8,627"
How does bought at 82% sell at 81% lose 100%? If there is 5% fees then it might be 6% rather than 1%.
(Spend 8627 mana order is for a position of 10561. This 10561 position cost Johnathan at 82% if no fees M 8660, When he sells 10561 at 0.81 this costs him M 2006.59 so a total cost of M 10666.61 but the 10561 yes and no cancel for 10561 returned to him for a cost before fees of 105.61. With fees this might be 6 times more but it is still small compared to the 10k profit.)
I see what you're saying now. Yeah, I agree with you that if someone has already taken a no position, they should protect it by placing a limit order on yes at 81%. That's a no brainer. However, if you don't have a position now, it's risky to place a limit order on "no" with the plan to then place an offsetting order at 81% if it gets filled. I haven't really focused on @JonathanRay 's incentives regarding "no" orders in the book, because there isn't much volume posted there. But if people started placing a bunch of "no" limit orders at 83% -- without simultaneously placing "yes" limit orders at 81% -- it makes it easier for him to profitably push the market to 80%. Before he pushes it to 80%, he should execute against any "no" orders that are posted (basically take the market all the way up to 99% to clear out any "no" volume completely), then immediately execute against the "yes" orders posted at 81%, push the market briefly to 80%, then resolve it. At the moment he decides to push the market to 80%, the "no" orders are free money for him.
(While letting a "no" order sit out there is risky, letting a "yes" order sit out there at 81% is relatively safe because it's unlikely to see execution unless someone's trying to push the market to 80% (in which case it will resolve in your favor).)
There's two ways to think about why he loses 100% on Ṁ8,627:
1. The simple way I think about is, to clear "no" orders worth Ṁ8,627 from the order book, he has to place a "no" bet worth Ṁ8,627 -- and that bet becomes worthless when the market hits 80%. You just deduct that amount as losses from the potential max payout he gets on his "yes" bet.
2. You might be thinking of him executing against the "no" as him partially "offsetting" his yes position. Conceptually that's right but the math is harder: Remember that, at 81%, a "no" bet for one manna buys 4.26 times (= 0.81 / 0.19) as many shares as a "yes" bet does. So to place "no" bets worth Ṁ8,627 (which is what it takes to clear the order book), he has to sell 45,405 shares (=8627 / 0.19) when he only bought 67,933 "yes" shares to start with. So most of his yes position goes away. And his cost basis on the "yes" shares he originally bought looks to be a bit above 82%, so he only ends up with a profit of Ṁ3,582 on the 22,528 "yes" shares that he'd still own after pushing the market to 80%.
I editted in my calcs:
(Spend 8627 mana order is for a position of 10561. This 10561 position cost Johnathan at 82% if no fees M 8660, When he sells 10561 at 0.81 this costs him M 2006.59 so a total cost of M 10666.61 but the 10561 yes and no cancel for 10561 returned to him for a cost before fees of 105.61. With fees this might be 6 times more but it is still small compared to the 10k profit.)