Yeah I was thinking this the entire thread. Reminds me a friend who applied for what is probably a similar position and was booted from interview when he couldn't come up with 22^2 instantaneously in a string of rapid-fire questions.

In 5 seconds I would have guessed 412 because I have 5 seconds. In round 2 I probably would have guessed some absurd number just to **** the game up, because **** their silly games.

I still don't understand the random hate/criticism in this thread. Basically they host 'silly games' because it works and makes money (the people they hired have made a lot of money). I understand if you think "f that", but what is so scary or even surprising?

round 1 everyone enters their own number
round 2 everyone enters the number they get back after round 1

you should just give back the number you get. unless you messed up on one round, then you have to readjust.

I think the objective of the exercise is a key piece of info.

If you're aiming to outperform your rivals, rather than estimate as closely as possible, and you know the number of rounds, there may be a case for ridiculous first round bid (e.g. 10^10). Your rivals will adjust their bids upwards in the next round and later rounds. We keep our bid constant until the final round, when we move to a sensible estimate based on the first round.

Of course, if multiple people adopt this strategy, then you don't actually gain any information on any of the rounds.

Not sure why some assume you were given any info about the length of the game or the final goal because from the OP timeline it can't be assumed you know that at the start.

yep, things happened just as described in the OP. I still do not know for sure what the goal was for this game if any, what constituted a "win" if anything did, or exactly what they were looking for

Overall this was a little unsettling, as I was doing stuff like this for hours, with usually no feedback at all as to how I was going

What would your approach be if we adjust the original problem a little bit:

Positive integers are picked at random from an interval: 1 to 1000 inclusive. The picked numbers are not necessarily distinct.
There are only 2 guessing rounds and you have to minimize your absolute total bias over the 2 guessing rounds. You don't compete with the others.
You may assume everybody uses an optimal strategy (or uses your own strategy).

My approach:

If not optimal it's at least superior to "pick your own number as your first guess".

Welcome back from the cryogenic freeze. Google financial meltdown of 2008, you have a lot to catch up with.

So if I mention this company never has had a losing year, and 2008 was their most profitable year, you guys would think more highly of their hiring process?

I would ask everybody what number they has. When asked why ... I would say:
- Teamwork communication is key to business success
- I have to work with facts. I can't guess - especially if I'm on company time.

Or I would guess 333 because that's the most commonly guessed number between 1 thru 1000.

Cyberfish, I agree with your guesses

mode your number
mean your number x 1.5
median 500

5 second thingy... 6215 - a little as we have a smaller number

I understand people are examining this in detail to come up with the actual best strategy, but OP had about 5 seconds to guess. That changes things significantly.

1st 5K, 2nd 100K, 3rd 25k or 20k.

412 is a low apeareance number so I'd guess way higher. 100K was to try to crack the game. 3rd would be the guess, probably 20K in 5s.

yeah def
I think you are on the right track but stopped early.

Essentially what you are saying is that we should try to beat the others in this game. But they will also try to beat us. In essence it's war of information. Or a game of incomplete information. So we should try to find GTO response.

No idea how a GTO would like here, but I think it would be pretty close to just throwing out random numbers.

The best chance at hitting the true average is, I think, just repeating what is thrown at us. But in a context of beating the others such a strategy is exploitable (by what you suggested for instance).


Anyway I think what they might be testing is how hard you are trying to beat others at a war of information. Maybe you could say the guesses of the participants form a sort of market, and the numbers the game host tells you is the market average. Now your next guess is trying to beat the market (the other players). As a believer in efficient markets you should just adhere to the strat of returning the last number you got. Just an idea, might be totally off.