Violating the EMH - Prediction Markets

post by sapphire (deluks917) · 2021-03-28T04:05:09.184Z · LW · GW · 9 comments

Contents

  Real Money Prediction Markets
      Bonds:
      Real Bets:
      Research:
      Disclaimer:
  Norms
        Update: BSC Launcher seems scammy. I was afraid I got rug pulled. I sold my shares. 
    Notes:
None
9 comments

I am quite confident of the information in my post on the Efficient Market Hypothesis [LW · GW]. I am happy to bet on the perp-future trade I described there since it is very safe. I will even give people betting against me a big advantage (1). However today I want to share some thoughts on prediction markets and group epistemics. Unlike the safe bets in my previous post, you can very easily lose money betting on prediction markets. I am including some good live trades.

Real Money Prediction Markets

There are basically two classes of prediction market bets: 'bonds' and 'actual bets'. 

Bonds:

Bonds are bets that are close to 100% to payout. Currently, you can get ~3% returns in a month betting that Joe Biden will still be president on April 30th on Polymarket. These bets have been plentiful recently because of  'Maga money' but sometimes non-Maga examples come up. Obviously, this bet is not riskless, especially given there is some counterparty risk from Polymarket going under (though polymarket is supposedly non-custodial so there might be ways to recover funds). However, it is a relatively straightforward bet with quite good returns and might be attractive to some people.  

Real Bets:

I should give an example of a live bet I think is good. I would suggest betting on the Olympics on FTX (not available in the USA). https://ftx.com/trade/OLY2021. The best argument is that FTX's probability seems way out of line with other sources: The Good Judgement Project has it as a 98% probability. I find the GJP relatively credible even though I think 98% is too high. Polymarket (2) and Betfair have it at ~92% and 86% probability respectively, although the volume is thin. FTX has it at ~76% with much more volume. However, you can find more considerations in this Twitter thread. The relay has already started. Last time the Olympics was postponed right before the relay began.

Research:

The same rationalist who made ~100K betting on Biden post-election made five figures betting on the above CO2 Market. He noticed that you could access the resolution data early if you changed a url parameter. Given this, he was able to tell the market was going to resolve 'No'  before anyone else and make a large amount of money. It is possible such edges will be available in the future if you are clever. 

Disclaimer:

Sometimes prices are pretty fair and polymarket has significant fees. At many points the binary options (ie https://polymarket.com/market/will-bitcoin-btc-be-above-55k-on-april-1-2021) have been priced in line with my basic Black Scholes model. Sometimes they have been way off. Always remember that you might be the dumb money. I would only place a bet if I thought my edge was fairly large.

Norms

I have spent a decent amount of time talking about bets and investments with friends. It is very good rationality practice. Some advice that assumes you are working collaboratively to make money:

-- Tell your group before you buy in if possible. It is very common to wait until your investments go up before you say anything. This is easily explained by signaling theory. If you tell people before you buy then risk looking bad if your investments lose value. But waiting to tell your friends can cost them huge amounts of money. Many times my friends have only talked about their ideas after the asset DOUBLED in price. If an idea good enough to buy in it's good enough to let your group know. (Example: I will be buying BSC Launcher tomorrow unless the price immediately gets out of control)

Update: BSC Launcher seems scammy. I was afraid I got rug pulled. I sold my shares. 

-- If you are buying a volatile asset it is important to have a plan. I like "sell 20% every time the asset doubles, don't sell on the downsize". If you initially bought a very small amount you should just hold until the dollar amount is 'meaningful' to you before you start selling. Conversely, you can choose an ambitious target price at which to sell most of your holdings. I always like keeping at least a decent amount of my winners. Dance with the girl who brought you. The main exception to this rule is if your 'investment thesis' breaks. If Sam Bankman-Fried abandons Solana I am going to sell most of my holdings too. I hope I find out early! There are obvious tax implications to selling early so I suggest donating winners if selling would incur short-term taxes. 

-- Don't get "in too deep". A friend of mine likes to say 'Odin hung on Ygdrassil' but you aren't Odin. If you do something risky for too much money you will make bad decisions. People got rich holding crypto not selling on every downturn.

-- Either go big or small. If you hear about a promising idea with a big upside it's often worthwhile to put in a little and HODL. If it works out and you passed you will feel pretty dumb. But if you are not willing to go big I would not go medium unless you have a genuinely systematic approach (which many gamblers do). If you don't believe enough to go big just make a bet small enough it won't hurt if things go south. I will note you cannot really take this approach if you working in a hard domain like sports gambling. But I would suggest finding softer fields. In my experience sports betting (on normal sports) is a very hard way to make money. 

 

Notes:

(1) - If you want to bet against the 'perp-future arbitrage trade I will give you 2-1 advantage (1). I will make an easily auditable account. For every dollar the account losses in the next six months I owe you 2n dollars. For every dollar, the account gains you owe me n dollars (we can agree on n over pm). If you do that trade correctly you won't lose more than a few percent and you can easily make more. The best way to set this up would be to get a trusted third party (outside the USA) who has an FTX account to set up a subaccount, deposit 10K, and give me access to the subaccount (they can disable withdrawals). In six months we can check how the account is doing. If FTX losses the funds I lose everything. I would be happy to do this for any reasonable size.
 

(2) - Polymarket bet is not quite the same since it resolves on May 1st. However, it would be pretty unusual for the Olympics to get canceled/postponed on only a few months' notice. Despite these differences, there is an obvious amount of 'semi-arbitrage' you can do between FTX and poly/Betfair. 

9 comments

Comments sorted by top scores.

comment by Chris van Merwijk (chrisvm) · 2021-04-08T07:35:39.452Z · LW(p) · GW(p)

Is there currently a way to pool money on the trades you're suggesting? In general it seems like there is some economies of scale to be gained by creating some kind of rationalist fund

Replies from: antanaclasis
comment by antanaclasis · 2021-04-12T02:42:03.529Z · LW(p) · GW(p)

On this note, I would definitely be willing to pay premium to be part of a fund run by a rationalist who’s more intimately involved with the crypto and prediction markets than I am, and would thereby be able to get significantly more edge than I currently can.

Replies from: sharpobject
comment by sharpobject · 2021-04-13T10:31:38.036Z · LW(p) · GW(p)

I should have this ready within 2 months.

Replies from: liam-donovan-1, bebbzor
comment by Liam Donovan (liam-donovan-1) · 2021-05-01T03:37:20.925Z · LW(p) · GW(p)

What prediction markets are liquid enough for a fund to make sense? (unless you were just referring to the crypto trading part?)

Replies from: sharpobject
comment by sharpobject · 2021-08-29T14:48:51.291Z · LW(p) · GW(p)

Hello,

You're right. Most of the time there are not many prediction markets I would want to take positions in that also have enough capacity. We would be willing to regularly bet millions of dollars on coups not occuring in the US, but these opportunities are unusual.

comment by bebbzor · 2021-04-27T07:31:56.663Z · LW(p) · GW(p)

Where do I sign up?

Replies from: sharpobject
comment by sharpobject · 2021-08-29T14:47:23.693Z · LW(p) · GW(p)

Hello, 

This is just about ready.  You can reach me at robert.burke@ltcm.lol about this.

comment by Liam Donovan (liam-donovan-1) · 2021-04-02T07:08:13.572Z · LW(p) · GW(p)

Just commenting to say this is a great post and I'm surprised it hasn't gotten more engagement (maybe it's so good there's nothing else to say)

comment by sapphire (deluks917) · 2021-06-07T22:28:12.187Z · LW(p) · GW(p)

Update. I exited the oly bet 2 days ago at 87.2 percent. Just letting people know.