Let's get rid of ties... ARMAGEDDON PLAYOFF MATCHES

[Caleb_Sykes]

Since I didn’t spend all season updating giant excel books, my mind is now free to suggest insane ideas for events. So, here is the newest insanity: armageddon playoff matches.

I’ve been seeing comments about playoff ties and how problematic they are. When teams score the same number of points, there are 3 options currently:

1. Declare the match a tie and continue on with the tournament (currently used in quals, but not possible with current playoff structure)
2. Replay the match (currently used for finals matches, but causes huge delays in certain years if applied to entire playoff bracket, see 2017)
3. Use tiebreaker criteria (currently used for non-finals playoff matches, could be used in all playoff matches but very commonly ends up feeling bad/arbitrary for the winners, losers, and spectators)

Instead of any of these, I’m suggesting the use of armageddon matches in the finals. The terminology comes from armageddon chess, in which the player with the black pieces “wins” the match if there is a draw, but as compensation, the player with the white pieces receives more time on the clock in order to force a straight win.

For FRC, here is how I would propose implementing armageddon matches:
When teams submit their lineup for playoffs, they can also submit a “time bid” of how much time they would be willing to give up in order to have draw odds (that is, advance on a tied score). The alliance with the higher bid will receive draw odds for the match. If they do not submit a bid, their time bid is defaulted to 0.50 seconds. Time bids must be between 0 and 99.99 seconds and will be entered up to the first 2 decimal places. If the time bids for both alliances are even to the first two decimal places, the red alliance receives the draw odds as well as the associated time loss.

During the match, after the autonomous period, the alliance with the lower bid will be enabled immediately (as all alliances currently are). The FMS will then wait the amount of time specified by the winning (higher) time bid before enabling the other alliance. I think the start of teleop is the best time to remove time. Removing time from auto will interfere with auto routines, removing time in the middle of teleop will effectively incapacitate robots for far longer than the “bid” time, and removing time at the end of teleop will knock out robots during the most exciting and most critical moments of the match (while also taking away from a potentially finite period of time for endgame tasks).

For audience clarity, there are a few approaches that could be used to show what is happening:

1. The alliance with draw odds receives 0.5 points upon winning the bid. This avoids any ambiguity about ties, but will look pretty strange to the audience and make data processing/display more difficult. It is viable but I am partial to other options.
2. In game design, make all game scoring opportunities (including penalties) a multiple of 2, then provide the alliance with draw odds a 1 point bonus on winning the bid. This is functionally the same as option 1, but gets rid of annoying half points at the expense of inflated total scores. I prefer this to option 1.
3. Instead of messing with the score, just put an indication on the final scoreboard that says “red advances on tie score” or “blue advances on tie score”. This could also be added to the in-game scoreboard but I’d be worried about clutter. Someone with better graphic design skills than me might be able to figure something out. I like this about as much as option 2.

I think a bidding process would add a fun little strategy component to the game and will make advancement feel more fair as teams will be responsible for their own destiny regarding ties. Don’t want to deal with any time lost? You can bid 0.00 seconds and will be active the whole match. Hate losing on a tied score? You can bid 30 seconds to make sure it doesn’t happen. The bidding process will guarantee both teams get what they want, and not one single person on any team will ever feel any heartache about losing due to a tied score ever again

Fun poll, how spicy would this rule change be?

[Richard_Wallace]

Caleb, how would you have felt about simple sudden death this year, as an alternative to supercharging? Sudden death means the match ends when one alliance has filled its grid and engaged three robots on the CS, or when the timer reaches zero, whichever occurs first. No referee action needed other than keeping up with the grid score, which they will generally have time to complete while the robots are balancing. The CS balance sensor ends the match.

[AndrewPranger]

I like the way you think, but this simply isn’t chaotic enough. I feel we can integrate more design from the 2018 FIRST game, Power Up. When the Grid is full, alliances can triple engage on the charge station in order to activate a time based scoring of +1 point per second. However, we remove the final 30 second safety zone associated with the charge station, allowing the other alliance to come try and move you off of your own charge station.

[Richard_Wallace]

@AndrewPranger

simply isn’t chaotic enough

I was strangely attracted to chaos as a graduate student. Then I started working in manufacturing, where people like predictable results.

[Caleb_Sykes]

Sounds fun in principle, there’s some weird edge cases I don’t like but would they have been weirder edge cases than the ones we got with supercharging

The big one I think of would be intentionally taking 5 fouls (25 points) near the end of a match (holding multiple game pieces maybe to speed up finishing the grid) because you can make up 30 points by balancing and ending the match.

[Lil_Lavery]

Just bring back the 2001 match timer multipiers

[alst]

@Caleb_Sykes

• Replay the match (currently used for finals matches, but causes huge delays in certain years if applied to entire playoff bracket, see 2017)

To be fair, 2017 did have effectively the same tiebreaker rules as 2023 (for anyone unfamiliar, they were last overhauled after 2016), but discrete scoring led to a lot of unbreakable ties anyway, particular in early weeks where fuel wasn’t being scored.

@Caleb_Sykes

If the time bids for both alliances are even to the first two decimal places, the red alliance receives the draw odds as well as the associated time loss.

So if I’m understanding this correctly: a lack of action by both alliances would mean that red always wins in the event of a tie?

[BrendanManning]

@alst

a lack of action by both alliances would mean that red always wins in the event of a tie?

Red would also lose 1/2 a second at the start of teleop. If, that seems like an unfair advantage to you, you can bid higher.

[Ekcrbe]

I don’t really like it. I can’t quite put my finger on why, but breaking ties in any way other than one alliance scoring more points than the other doesn’t sit right with me.

If you want chaos, I propose overtime. For any playoff match where the winner is in doubt at T=0, don’t move the robots. Scored game pieces are removed from the grids, but the robots and unscored game pieces on the field stay where they are. If a match is tied, the robots are re-enabled for a 30-second teleop-only overtime period from the same positions where they ended the match. Regular endgame rules apply, so you can get credit for balancing the charge station again in overtime. The alliance that leads after overtime advances.

[IanSzalai]

I like the idea of overtime more as well, but it doesn’t really work with FRC. Endgames are often a one way trip with climbs, even this year some teams probably don’t expect to get off the charge station after forking. It would also cause huge delays if teams have to wait for scores to be announced before getting their robots and driver stations off the field.

[AndrewFRC135]

That’s a neat idea. Yeah, the edge cases and audience communication are challenges, but I see what you’re getting at. A whole new layer of strategy.

Personally, I like the idea that when there is a tie score in playoffs, the alliance that scored their last point(s) the soonest wins tie break. An alliance that triple engages with 8 seconds left would beat one that throws a cube on/parks with 4 second left.

For real-time scored objectives like the charge station, this is easy, but all others would require video review with conclusive evidence. Fall back to the tie break order if video is inconclusive. I mean, I think it’s about time official video review begins to play a small part in our elimination matches anyway

[s3529]

Except for removing the safety zone, I like the sound of this.

[cgmv123]

“Bidding time” is too complicated. Just make the tiebreaker whoever achieves the endgame (i.e. balances the CHARGE STATION) first as long as that can be determined by sensors.

[petercooperjr]

I always enjoy reading another of Caleb’s Crazy Ideas. I feel like this one is going about everything the wrong way, though.

This is trying to introduce a new mechanic, full of interesting game theory and strategy, sure, but trades off one alliance always playing with a handicap in exchange for a clearer outcome for the (hopefully) rare case of getting a tie.

I think the real problem is mainly around how ties are presented and perceived, that one side won “on a technicality” or something that wasn’t a “real” point. But really, that’s just because it’s shown as a tied score and shown and thought of as a “tiebreaker”.

Whereas, if instead of having tiebreakers at all, the scoring just included the relative worth directly, then it would feel a lot more natural to everyone. Like, if in playoffs (and maybe even in quals too), Table 6-2 & 6-3 for Charged Up just listed the scores like this:

Award	AUTO	TELEOP
Mobility	3.000003
Game Pieces: Bottom	3.000003	2
Game Pieces: Middle	4.000004	3
Game Pieces: Top	5.000005	4
Link		5
Docked & Not Engaged	8.000808	6.0006
Docked & Engaged	12.001212	10.0012
Park		2
Opponent Foul		5
Opponent Tech Foul		12.12

That is, what was the tech foul tiebreaker changes to be the first two places after the decimal, then the next two decimals are the Charge Station score, and then the next two decimals are the Auto score.

You end up with the same scoring outcomes, just calling it an actual “point” (or fraction thereof) rather than a “tiebreaker” I think may make people think of them differently. Feel free to the scale the numbers however you want if you don’t like having that many decimal places, sure.

But really, my absurd example here is actually more just pointing to that what you need is for the game design to not involve ties at all, but rather just incorporate the desired tiebreakers into the scoring, such that the only way one can get a tie is for it to truly be a tie where both alliances scored the same objectives, at least within what the game design traditionally wants to call equivalent for purposes of tiebreakers.