Computing max games for knockout tournament?

I am organizing a 3 strikes knockout tournament (using brackelope) and getting ready to submit to ifpa website. Does anyone know an easy formula for computing the max games for a given number of players?

I do! :smile:


perfect! thanks!

Correct me if I’m wrong, @iscrz, but my understanding is that when you email the results from Brackelope, the email has a .txt attachment that includes the TGP in the form of “meaningful matches played” so you don’t even need to figure it out. Here’s an example from the most recent Portland tournament:

Adjusted Results for the World Pinball Player Rankings
Upload URL:
Name: Clinton Street 5/26
Date: 2015-05-26
Format: Double Elim. Bracket
Meaningful Matches Played: 12
(Multiply by games per match. See

1 Like

CFFLegs–this issue is estimating max games BEFORE the tournament happens, when you are submitting to the calendar

Ah, gotcha. Since it doesn’t affect the point totals in the end, I just always go with the same number of estimated players so I know the corresponding TGP for whatever format to enter for all calendar submissions (I always guess 30 players will show up, which means 11 TGP for single game/double elimination and 17 TGP for single game/three-strike knockout, the two most common formats for our weekly tournaments in Portland). That way I don’t have to think too hard :smile:

floor(log2(players - 1.0)) x strikes + (2 x strikes) - 1


nerrrrrrrrrrrrrrrrrrrrrrd. (did i add enough "r"s for it to approve my post?)


nice! what is the floor parameter? or is there an “=” missing?

Edit: yep, just ignore the “floor” part. checked the formula against pinwizj’s sheet–cool!

floor(x) is the largest integer that’s not larger than x.

mmmm, never used that function before.

floor(x) and ceil(x) are common programing functions that round decimals down and up, respectively

There usually a round(x) function too but I’ve never seen a round() function I could trust.

Mathematicians can’t even agree on what rounding should do in some special cases, so it’s not surprising that one working on floats has some dodgy interactions, especially with decimal numbers.

Floor and ceil aren’t just for programming: they come up in more abstract maths fairly often, like @bkerins’s formula above or in various infinite limits.

Does 4.49999999… round to 5? :smiley:

Depends what you’re doing with it :smile:

(For data analysis, it’s often correct to round terminal 5s to the nearest even digit to avoid bias.)