No firm methodology at this time. I won't bet until / unless there is one; it might be kinda vibesy.
Buckets are left inclusive, right exclusive, so a price of exactly 160 M/$ resolves to the 160-200 bucket. If there's uncertainty I will default to averaging out and taking the central estimate as best I can.
I will consider things like:
Mana purchase price available in the store, both current and recent sales
Apparent mana purchase patterns (did a sale increase sales?)
Cost of features available for either dollars or mana (for example, if you can buy a subscription with either currency)
Implied exchange rate from charity giveaways
The above list is not intended to be exhaustive.
Some recent reference points:
Black friday sale had M120000/$750 available, for a rate of M160/$. This didn't seem to generate a huge sales spike.
Charity currently lets you buy tickets at 254/M100, with a pool of 2.9e6 sold and $1000 for the giveaway. So the current implied price is M1153/$.
1,000People are also trading
For a slightly shorter timeline (3-4 months) my rough outline is:
Mana shop will be funded almost entirely through existing balances not paid mana
There will be a second charity giveaway and it will blow past 250/$ again and probably close to 500/$
Until the mana shop and charity start draining existing balances (which is not a guarantee), the mana:USD rate will remain well above 100.
@Eliza I would hope that "investing in markets" also has additional potential. More good markets, better algorithmic discovery, and better limit order UI/UX might all help there. Hopefully we're not saturated on using mana for predictions! (Though I agree that at present, without changes, it looks that way.)
@EvanDaniel Someone should study Mikhail Tal's series of markets on lithium production (and others) to form an argument for [something about how mana can convert to USD]. I think there's enough there (markets, timeframe) at this point to come up with something.