# Reducing Tiebreakers in PAPA-style Playoffs

With more events running into time issues, I thought I should resurrect an idea I had many years ago for reducing the frequency of tiebreakers. Both the 4-2-1-0 and 3-2-1-0 formats lend themselves to 2 different kinds of ties. The first is when two or more players have identical records for all three games, e.g. each has a 1st, a 3rd and a 4th. The second kind is when players have different records but due to the point values come out equal. In the 4-2-1-0 version, this could be when player A goes 1st-4th-4th and player B goes 4th-2nd-2nd, each receiving 4 points total. This version can even be 3-way with different records if player C goes 2nd-3rd-3rd to also get 4 points. (Player D would then have 9 with 3rd-1st-1st.) In the 3-2-1-0 version, this could be player E going 2nd-3rd-4th while player F goes 1st-4th-4th, both ending up at 3 points. With a minor change in point values, those second types of ties with different records but the same point values can be eliminated.

How? Use the following points: 16-9-4-0. This gives slightly more weight to being one-better-plus-one-worse vs. two in-the-middles, e.g. 1st + 3rd beats two seconds 20-18; 2nd + 4th beats two thirds 9 to 8; 2nd + 3rd + 3rd beats 1st + 4th + 4th by 17 to 16 (makes sense, the first player finished ahead of the second one two times out of three); 2nd + 2nd + 4th beats 4th + 4th + 1st by 18 to 16 (same two out of three logic; here, you beat them by 4 â€śtotal placesâ€ť in the two 2nds vs. 4ths and they only beat you by 3 places in the one 1st vs. 4th). In a three-game block, you eliminate 2/3 of the tiebreakers this way. The scheme still gives more weight to winning (the 1+3 beats 2+2), but tempers it just enough to avoid ties unless records are identical.

With many tournaments now having 3 (12 or 16 players), 4 (24 or 32) or even 5 (40+ players) rounds of playoffs, thereâ€™s opportunity to shave a good amount of time off a long event this way. Having just run an event with tiebreakers and that ran late more than one night, I think itâ€™s time to try this.

3 Likes

I love it! My only concern would beâ€¦ would it make the overall tournament run faster if all i takes is one group to run into one of the other types of â€śunavoidableâ€ť ties? What if we used your new scoring system and then on top of that tournament directors could use the Nuclear Option: Euro-style qualifying-based automatic tie breakers!

1st 3rd is already better than 2nd 2nd in 4,2,1,0.

If thereâ€™s a tie between 2 players in a 3 game tie, I prefer using a straight head to head count back match up to split them - this is then guaranteed to produce a result without the need for additional calculations.

Edit: added the count back to clarify that no further games need be played.

2 Likes

In a nutshell the proposed system is trying to increase the value of 2nd place over the current PAPA scoring system.

Not that it solves the tie problem, but there is some psychological aspect to the scoring setups as well.

4-2-1-0 already feels painful when you are the player with the goose egg after game one. (It doesnâ€™t matter who the fantasy 4 super-elite group is, somebody is getting a zero on game one). Linear systems such as 3-2-1-0 or 4-3-2-1 donâ€™t seem as bad when you are trying to catch up, but yes, they offer more chances for ties across 3 games.

16 points for 1st and 0 for 4th just seems worse, although the 4 values are all scaled by a factor of 4 with the exception of the second place finish which was increased.

If time is the factor in larger tourneys, (or any tourney) there are certainly some easier tiebreak solutions, such as using the qualifying seeding to remove the tiebreak. Please note, I am not saying that is a better solution than adjusting the values or even a good solution, but it is faster than having to play another game and everyone would know that their qualifying position could have more implications other than driving the bus.

Bring back 10-5-1-0 with the tiebreaker being raw score across all machines

3 Likes

Signal boosting this. If you want to eliminate ties, do this.

At the same time, I donâ€™t see tiebreakers as something I want to eliminate. They are super exciting to watch, they make for good pinball.

5 Likes

1 Like

This is what we do for non-bubble tiebreakers for the IFPA WC. We copied that from the BIG10 conference tie breaker procedures.

While we enjoy doing that for non-bubble tiebreakers, weâ€™ve never supported anything besides a playoff for bubble ties/elimination situations. There is no tiebreaking procedure available that wonâ€™t have the player eliminated not feeling like absolute shit. We build those tiebreaker rounds into our schedule and would never consider changing that.

2 Likes

games have to big of an scoreing gap for that to work.

Just to confirm my understanding, this method values the last game played (3rd game of the 3) as having more weight than the 2nd or 1st. Outside of the rare tie in scores on game 3, whoever has performed better between the tied players on game 3 is the player that advances, correct?

Interesting concept for eliminating tiebreakers. Though I tend to side with @pinwizj of never wanting such tiebreaker to be used on whether players get eliminated.

Perhaps itâ€™s a good idea for a small local-level competitionâ€¦ but not as appropriate for larger events with more prestige (and typically more cost/investment of players to participate)

That was a joke because thatâ€™s how it used to be done.

Paul Madison won PAPA5 on a tiebreaker over Lyman Sheats based on a raw score tiebreaker

5 Likes

An equivalent way to view my scheme is 3-2-1-0 scoring but with the tiebreaker for non-identical 3-game records being most #1â€™s, then most #2s, e.g. 1st, 3rd and 4th beats 2nd, 2nd and 4th. Players with identical records are the only ones who still play a tiebreaker.

Josh, re the â€śtotal scoreâ€ť, or PAPA 6 in Las Vegas with Starship Troopers vs. Scared Stiff and Junkyard?

1 Like

They did â€śStarship Troopers divided by 4â€ť in that group . . . for reasons my dad still canâ€™t explain

1 Like

imagine if the games were Attack From Mars, NBA Fastbreak, High Hand, and Meteor

1 Like

I agree completely with those that believe elimination and significant bubble tiebreakers should be run as a match and not be broken via previous match records. Tiebreakers for seeding, sure, but if Iâ€™m at risk of being eliminated in the current round I want to be measured directly against my opponents, not how they (or I) performed in previous rounds against different players.

4 Likes

so you would be in favor to get same round tie breaker based on position in this round, 4/1/0 winning tie breaker against a 2/2/1 for example (most number of 1st, then 2ndâ€¦)?

Yes, I would prefer running the tiebreaker. If you donâ€™t, the points system loses relevance.

2 Likes

Iâ€™m really not a fan of this kind of idea, or the way Pinburgh did it either.

5 points is 5 points, and it shouldnâ€™t matter what combination of finishing positions you had to arrive at that total. Calling 4/1/0 â€śbetterâ€ť than 2/2/1 is completely subjective to how you look at those two performances. Avoiding a last place could be considered a better performance than getting a win. I know I would be salty if I had the 2/2/1, beating the 4/1/0 player on 2 of the 3 games, but they advanced over me because greater importance was placed on their lone 4.

If time is expected to be an issue, I think itâ€™s fine to limit tiebreaker machine selections to shorter playing games, but ties to advance should always be played off. Itâ€™s exciting and itâ€™s fair. If time is really that much of a factor, itâ€™s worth looking closely at the format to figure out why. Too many concurrent events happening? Pushing the format too far to reach 100% TGP? Starting too late in the day? Those are the issues that should be examined and tweaked before adding weird rules to nix tiebreaker games.

Edit to add: I like the concept of 16-9-4-0 to reduce the possibility for ties, but when they do occur, play them out.

4 Likes

I wrote a script years ago for determining the proportion of possible outcomes that would result in ties for various scoring methods. Ran Bob scoring through it and here are the results. Looks like it does indeed result in pretty substantial reductions in ties.

Note: I quickly wrote this many years ago, so no guarantees on the results.

[1] â€śScoring Method: 4-2-1-0â€ť
[1] â€śTotal permutations: 13824â€ť
[1] â€śtotal ties: 0.42â€ť
[1] â€śfirst ties: 0.12â€ť
[1] â€śfirst ties, no break: 0.1â€ť
[1] â€śfirst ties, three-way: 0.02â€ť
[1] â€śsecond ties: 0.18â€ť

[1] â€śScoring Method: 3-2-1-0â€ť
[1] â€śTotal permutations: 13824â€ť
[1] â€śtotal ties: 0.44â€ť
[1] â€śfirst ties: 0.18â€ť
[1] â€śfirst ties, no break: 0.16â€ť
[1] â€śfirst ties, three-way: 0.03â€ť
[1] â€śsecond ties: 0.17â€ť

[1] â€śScoring Method: 16-9-4-0â€ť
[1] â€śTotal permutations: 13824â€ť
[1] â€śtotal ties: 0.14â€ť
[1] â€śfirst ties: 0.06â€ť
[1] â€śfirst ties, no break: 0.05â€ť
[1] â€śfirst ties, three-way: 0.01â€ť
[1] â€śsecond ties: 0.04â€ť

5 Likes