adrian

So far it is still 70, so 1 planet.

There was some mail from some expert to a fellow forecaster, but it doesn't seem to me that person has thought about the question deeply. I will not update on it, since I was already above the crowd forecast and trust my analysis.

I expect the next planet to be added around 21.3. + 49 days = 9th of May (+- 145 days ≈ 5 months).

Iran is attacking Israel + Trump likely being elected in the fall. What can go wrong with that?

Up to 19th of April the number is now at 370, which is a bit higher than the point estimate I predicted, but it's still well within the range. While the change does adjust the upper limit of the end-of-may estimate, I'm confident it won't even come close to 500 (probably something like 405 +- 35).

Time progression

From the description: "This question will be resolved as “Yes” if the Habitable Worlds Catalog adds 5 or more new potentially habitable worlds to the list (i.e., the sum of their conservative and optimistic samples) discovered on or before 31 December 2024." It doesn't say they need to be discovered after 1.2.2024, but only added to the list until 31.1.25 (so 12 months in total). Technically, it might be that there are planets discovered after 1.1.25 and added before 1.2.25, but that seems very unlikely given that even the most recent addition was discovered last year, so I'm gonna ignore that for now.

For the base rate, last year was apparently 5 planets. At least 2 more planets have been added in January (two on 25.1., to be precise), so I take it as 7 planets over 13 months or 0.54 planets per month. After 1.5 months another planet has been added, giving 0.67 per month, or 0.55 for 8 in 14.5 months. I also note that (at least) 3 planets until now gives 1.14 planets per month. Assuming (conservatively) the three this year were outliers in frequency, I get 0.54*12 = 6.5 planets for the questioned period. For an upper bound, take the rate of this year so far, I get 13.7 planets.

Not enough! For some more base rates, it was 2 planets on 5.12.2011, giving until 1.2.2024 a base rate of about 67/146 ≈ 0.46 planets per month, which would be 5.52 planets for the time frame in question. One year later 5 new have been added, indicating a surprisingly stable the base rate. Interestingly, both originally "confirmed" (Gliese 581d and HD 85512b) and (I think) several "candidates" are not in the lists anymore, so they also remove planets, making the rate effectively higher.

And finally, I don't expect the rate to decrease. New equipment like JWST is available, more precise, and preliminary research improves, too.

Taking everything above into account and summoning my magic, my crystal ball says a 90%-CI of 3 to 15 for this year, though left-skewed with mean at ~7, so one every 52 days. That matches the rate of 1 in 49 days so far. Modelling that as Poisson(λ=7) (which says 3-15 is 97%), I get P(x<5) = 17.3%. To account for the wider tails, I round to 20%.

For future reference, a simple linear interpolation of this prediction gives the following numbers for the rest of the year:

• 1st of April (60 days): 1.15 (0.49 to 2.47)
• 1st of May (90 days): 1.73 (0.74 to 3.7)
• 1st of June (121 days): 2.32 (0.99 to 4.97)
• 1st of July (151 days): 2.9 (1.24 to 6.21)
• 1st of August (182 days): 3.49 (1.5 to 7.48)
• 1st of September (213 days): 4.08 (1.75 to 8.75)
• 1st of October (243 days): 4.66 (2.0 to 9.99)
• 1st of November (274 days): 5.25 (2.25 to 11.26)
• 1st of December (304 days): 5.83 (2.5 to 12.49)
• 1st of January (335 days): 6.42 (2.75 to 13.77)
• 1st of February (366 days): 7.02 (3.01 to 15.04)
  

Generated using

from datetime import date

def f(d):
    d -= 1  # e.g. 4 = April
    start = date(2024, 2, 1)
    end = date(2024 + (d//12), d%12+1, 1)
    d = (end - start).days
    m = d/(365/7)    # mean
    l = d/(365/3)    # lower bound
    u = d/(365/15)   # upper bound
    print(f'1st of {end.strftime("%B")} ({d} days): {r(m,2)} ({r(l,2)} to {r(u,2)})')

Confirmed previous forecast

• -1% for time progression
• +1% because I'm surprised how high the still consensus is, I might have missed something (it's also going down, though)