Modelling the price of Futures



  • I'm pretty certain I can build a pricing model to forecast the correct price of any future on the index. I am looking for someone with computer programming skills and maybe a data set to form a company and do this with. If anyone is interested please DM me. Apologies if I don't reply straight away as I am travelling over the next few days.



  • I think that something like this could be valuable.

    However, I don't think there's way of calculating the "correct" price, but rather an indicative price range.

    Let's take a non-British player in a non-PB league. How would you take into account the fact he had an English Mum? What if he'd expressed a hatred of learning foreign languages? Or desperately wanted to stay with his childhood club? I can't see how any model could pick that up in a way that didn't involve manual inputs, wasn't highly labour intensive, and didn't involve highly subjective assessments of an individual's psyche.

    Also, what data do we have on a player's future performance? We have data on past performances, but expectations of future performance are incredibly hard to predict in many circumstances. Particularly with young players and old players who develop differently. Also, how would you predict how tactical shifts would impact on PB? If you had listened closely to Klopp's pronouncements you might have recognised that Firmino was going to moved to a more central role, and therefore be more likely to be better at PB. And he has been. Would that factor into the model on predicting future performance? How could it? How about predicting managerial changes, such that at United recently?

    Also, would it factor in future potential dividend changes? Or changes in behaviour as a result of share splits? What ROI from dividends would the model be seeking? Would it assume that the share would be sold in 3 years, or that it could be sold before? Would it take into account changes in capital values, or just dividend returns?

    All of this is merely to say, such a model is a good idea, and one that has crossed my mind, but let's not say that any model can give a "correct" price. These things are necessarily fuzzy. It would take a next level machine learning type model to achieve anything close to what you're aiming at, and even then it would be pretty flawed.

    Regardless, I wish you luck! It'd be great to have another tool for traders to assess real value. :)



  • The correct price is based on the expected value of a player. Its not absolute. The bookies use similar models to price the odds on football matches. The prices you see are probabilities based on models that use historic data and tweeks for conditions on the day. For example if you see an over 2.5 under 2.5 goal market priced at 1.25 for over 2.5 and 5 for under 2.5 it means the bookies expect there to be a 80% chance of the game being over 2.5 (1/1.25=0.8) and a 20% chance of the game having under 2.5 (1/5 = 0.2). If the price is "Correct" on their over under market they would break even across a series of bets. The market could be incorrect in two ways, they can leave the odds too long which would cause them lose money or they can deliberately leave them too short. See below for the example.

    They offer punters a short incorrect price in order to make money. If they believed the probability was 80% O2.5 and 20% U2.5 they would offer 1.2 Over and 4 Under. 1/1.2 = 83% and 1/4 = 0.25% . This market is clearly priced wrong as the sum of the outcomes is 108%. In reality there is only a 100% chance the game will finish with either under or over 2.5 goals. The 8% is called the over round and it represents the bookies margin on the market they make. To figure out how much money a bookie takes on a market you need to know the volume and then multiply it by the over round. So if 1000 pound was traded on the market then with an 8% over round the bookies will take £80

    Football index is exactly the same. The maths required is a bit more lengthy and the data sets needed harder to come buy as no one has had to model the PB system before but this will be cracked quite easily by anyone with the data set and I would be surprised if people aren't already working on it. FI themselves must have a model. When you see FI drastically increasing the IPO prices but not increasing the dividend yield this is the mathematical equivalent of the bookies decreasing the odds on the O/U 2.5 market.

    the trading dynamic in football index is simply the ability to cash your bet out early. With a traditional bookmaker they will often offer you poor odds when you cash out. You will be happy that you got money but it truth if you were to be offered that exact same cashout over and over you would be financially better off not taking it. With FI the ultimate aim is for you to be able to cashout your bet with someone else on the market. FI don't have to worry about if person A offers person B good odds or not as they make 2% either way.

    They do have to worry about the IPO prices of the shares as they are liable for the dividends paid over the shares lifetime. The ideal market mathematically from their perspective would be to offer shares at price X knowing statistically it is likely over the shares lifetime the share will payout less than X. The recent hike in IPO prices without an increase in dividends ultimately results in a lower expected value (probably a negative one) for any one buying the shares. Also getting traders to trade shares at higher prices results in higher fees for FI as the 2% is higher in real terms.

    The correct price of a share from the buyers perspective is any share that has a possitive expected value over the life time of that share. If you buy these shares you will either A recvievr more in dividends than you purchased for over the lifetime of the share or B hold the share long enough to see market participants that don't understand correct value pay you a price that is too high for the share (so the buyer will have a negative expected value)

    there is more I will add to this thread later but I am rushing right now.



  • @NewUser interesting stuff. I suspected one of the main factors in the IPO hike was an attempt to take some of the bot-profit out. I remember seeing some complaints relatively recently (few months ago?) about it, and them seeming pretty unaware of displeasure at how it functions (strange I know). They indicated the system will be looked at when the beta site finally gets up and running properly - assume they have no better alternative/option at the moment than introduce higher (don't know how much this effects their bottom line of IPOs either way - must be a good source of income for them).



  • @NewUser I'm guessing you paid attention to maths at school?


Log in to reply