Hi,
I know there’s a separate post about thoughts on the Double Elimination format following Chezy Champs, but given the fact we (Orbit) just got a chance to compete in both Chezy and the Israeli Offseason – where we ran a 6 alliance (4 teams each) Round Robin format on a 33 team event, I thought I would share our observations/feelings and do a comprehensive comparison in hopes of maybe sparking a conversation about it being one more (possibly better) option for all competitions and not only Einstein.
I know FIRST hasn’t even put this option on the table at this point, but I believe it is truly worth talking about as a very viable option and that FIRST becomes more and more attentive to the community in recent years, so it’s worth the shot.
In order to focus the discussion and make a “quality” comparison between the various options, I went through the post on DE and most of my points will refer to the things that were raised there along with a few points of my own.
I will note that is for the Israeli offseason we elected to go with a pick order of: 1-6,6-1,6-1 and we ‘forced’ each alliance to include each of the 4 teams on their alliance in at least 2 RR matches + at least 1 of the 2 first finals matches for the finalist alliances – I will make the comparisons to this exact format based on the assumption that playoffs on all events except championship include 24 teams.
-
Event length and predictability:
Standard – 14-21 matches, 8 of which are known matchups and the rest are dependent on results. 2-4 required field timeouts (back-to-back matches).
DE – 15-16 matches, 4 of which are known matchups and the rest are dependent on results. 4-5 required field timeouts.
RR – 17-18 matches, 15 of which are known matchups and the finals are dependent on results. 2-3 required field timeouts. -
Number of matches per team:
Standard – 8 x 3 team alliances:
12 robots are playing 2-3 times, 6 robots are playing 4-6 times, 6 robots are playing 6-9 times
(2.33-3.5 average matches per team).
DE - 8 x 3 team alliances:
I got somewhat confused trying to analyze it, in general – 6 robots are playing 2 times, 6 are playing 3 times and there are some variations for what happens with the remaining 12 robots depending on results
(2.5-2.67 average matches per team).
RR – 6 x 4 team alliances:
6 robots are playing 2 times, 6 robots are playing 3 times, 6 robots are playing 5 times.
For finals – 2 robots will be playing 1 more match, 2 more will be playing 1-2 more matches and 4 will be playing 2-3 more matches
(2.83-3 average matches per team).
this is all based on the assumption that the 1st and 2nd pick are playing all matches which could change. -
Red card severity / margin of error / robot reliability issues:
Standard – you’re ‘allowed’ to lose once in each round without it causing you to be eliminated. If any of the robots on the alliance suffers reliability issues, your backup is ‘randomly’ selected and effectively eliminates the broken partner from the event.
DE – the first lose eliminates you from the Winners Bracket and makes it so every matchup from then on is do or die, the second lose causes elimination. If any of the robots on the alliance suffers reliability issues, your backup is ‘randomly’ selected and effectively eliminates the broken partner from the event.
RR – each lose badly effects your likelihood of qualifying for finals, but in most Einstein cases to this day at least one of the 2 finalists was 3-2. You have a 4-team alliance, so if you plan your matches smartly you should be able to work around most robot reliability issues and give teams enough time to fix them if they skip a match. -
Matchup difficulty for each alliance:
Standard – alliances 6-8 have a very low probability of making it out of quarter finals and ever playing a theoretically equal opponent (even the #3 vs #6 matchup is pretty lopsided since the 1st and 2nd picks for alliance #3 are the 5th+6th best in the event, while the #6 are 11th+12th). The only way for an underdog alliance (#5-#8) to encounter a favorable matchup is if both of them surprised a favorite alliance (#1-#4) on the same side of the bracket.
DE – apart from alliances #7 and #8, each alliance will be favorite in at least one of their matchups (unless they surprised their rivals in the first round), and even those two alliances will have a much more reasonable matchup in the first Lower Bracket game.
RR – each alliance will be a favorite in 6-x of their RR matches (where x is their alliance number). -
# of teams that stay relevant for each stage:
Standard – for each round half of the field is getting eliminated, so all 8 alliances for quarter finals, then 4 alliances for semis.
DE – for the first 2 round all teams are still in play, for the 3rd round 6 alliances remain, for the 4th there are 4 remaining and for the 5th there are 3 remaining.
RR – for the entirety of the RR phase all 6 alliances are still in play and are relevant to the outcome of the competition even if they have effectively been eliminated.
Finals are the same for each format – 2 alliances remain (except for the fact that in RR those 2 constitute of 8 teams instead of 6) -
The ability to quantify you’re results:
Standard – I made it to quarter finals / semis / finals.
DE – I made it to round number 1-5 / finals (you could in theory place yourself more precisely according to which round you were able to get to but it is generally confusing and unmarketable).
RR – there’s a clear rank for each of the 6 alliances at the end of the competition. -
Strategic diversity/flexibility:
Standard – you have at least 2 chances to figure out an opponent in each round of playoffs.
DE – apart from cases in which there’s a LB rematch of an WB matchup, there’s no opportunity to ‘solve’ an opponent. On the other hand, there’s a diversity in the matchups so you have a chance to try different things versus different opponents or give a certain strategy another shot against a different opponent.
RR – Apart from finals there’s no rematch of any matchup. On the other hand, each matchup is different and coupled with the fact you have a 4-team alliance there’s lots of creativity to be included. -
Match density and its relation to your results:
Standard – the more you win you better your chances of having more rest between each of your matchups (in case other matchups go 3 matches).
In general the density is relatively equal for everybody – 3 game rest between each of your first two QF matches and less than that on the third depending on the number of match 3’s, 1 game rest between each of the first two SF matchups and either 0 or 1 before a 3rd game if needed.
DE – the more you win the further apart your matches are. If you fall to the LB there’s little time to react and prepare between each of your matchups, there’s no real way of quantifying the density since it’s extremely dynamic and dependent on results.
RR – entirely predictable = your alliance plays every 3rd match in average and never a back-to-back. -
Average alliance quality:
Standard + DE – you get 8 alliances made up of the following theoretical robot quality:
Alliance 1: 1+2+24
Alliance 2: 3+4+23
Alliance 3: 5+6+22
Alliance 4: 7+8+21
Alliance 5: 9+10+20
Alliance 6: 11+12+19
Alliance 7: 13+14+18
Alliance 8: 15+16+17
RR - you get 6 alliances made up of the following theoretical robot quality:
Alliance 1: 1+2+18+24
Alliance 2: 3+4+17+23
Alliance 3: 5+6+16+22
Alliance 4: 7+8+15+21
Alliance 5: 9+10+14+20
Alliance 6: 11+12+13+19
** There’s an interesting dynamic where the #5 and #6 alliances’ first 3 robots are of pretty similar quality and in deeper competitions these alliances could be very solid with 3 mid-high level scorers that might be able to outplay some of the higher ranks in my opinion. -
Viewers experience – in event and on live streams:
Standard – pretty straight forward, the only thing you need to get a grasp on is who’s initially playing who and which side of the bracket each alliance is on. It’s also easy for the viewers to perceive how far each alliance got.
DE – maybe it’s the lack of familiarity and a habit that hasn’t been built in my brain yet, but to me it felt pretty chaotic, even if you understood the structure of the table, the number of possible variations and the order of the games are confusing and it creates a lack of clarity regarding the degree of success each participating alliance had - different alliances can end up with the same record but with a theoretically different level of achievement (see alliances 3 and 4 in Chezy Champs).
RR – the whole match layout is known the second the alliance selection is finished, the only thing that could get confusing is the placement table – you have to understand the tiebreakers and the remaining matchups each team has to be able to predict their expected final placement.
There is also a direct correlation between the number of wins and accomplished tasks within the matches and an alliance’s placement. -
Weight of different matches between two specific alliances:
Standard + RR - each game has the same weight to it and the direct record between two alliances will decide who will go how far (if you ask me that should be the first tie breaker in the RR format).
DE - there’s something odd to me about the fact that you can end the competition in a 1-1 draw with another alliance and have one qualify further than the other (see alliances #2 and #3 in Chezy Champs). -
Number of wins required to win the event:
Standard – 6 wins.
DE – 5 wins if you were on the WB until round 4 at least and 6 if you fell to the LB before that.
RR – 5-7 wins depending on the alliance’s RR record. -
District point system:
Standard – the method is well known at this point, could maybe use some improvement, but by and large it’s quite clear and logical.
DE - I suppose it is possible to do something similar to the standard one where you get a score depending on the number of victories and the stage you reached, but there is something a little confusing and challenging in the fact that if you go through 2 stages in WB, fewer victories are required to qualify for the finals and then you have to give a different score to the victories in these stages.
RR - if alliances consist of 4 teams, it is possible to give a score for each victory in which a robot was present on the field + a bonus depending on the position of the alliance at the end of the RR stage and add to that the score in the finals, similar to the standard today.
** BTW – if I recall correctly, the reward for passing stages today in the standard playoffs is 5 points for each victory, but you only get those accumulated as 10 if you qualify for next round and otherwise you get zero, I don’t understand why not give those 5 points to the alliances that lost 2-1 to reward them for being competitive and achieving the forementioned victory. -
World championship qualification criteria:
Regionals:
Standard + DE - the method is known, again – it may require improvement but it is known.
RR - if the alliances consist of 4 teams, you have to give some kind of alternative solution other than just give all 4 of them a ticket – unless the 4th replaces one of the awards ones – RAS/EI/CA (which I can’t say I’m entirely against given the fact that I think the program could be leaning more towards the competitive aspect).
Districts:
Standard + DE - the method is known.
RR – I think in DCMPs there shouldn’t be a problem with allocating one more of the tickets to the 4th winner.
The post is already super long as is but I would conclude by saying that in our personal experience the format of playoffs we held in the Israeli offseason of RR with 6 x 4 team alliances and the obligation of have all four teams participate in at least 2 matches + 1 in the finals, felt to me like one of the most flowing and pleasant experiences we experienced as a team to this day, and even if an alliance was down 0-4 going into their last match in the RR stage, I think they had something to play for in order to get the one victory and maybe improve their position to fifth. All the way to the end of the competition, there were 24 active and energetic pits and the whole crowd was involved in the action to the last moment, because even if the chances of some of them qualifying for finals are really slim, they still could have significant effect on the overall result.
Yes, there was less of a ‘sword hanging over your head’ feeling which is good in some aspects and bad in others, but at least to us the DE format felt messy and confusing and the fact that you can have a 1-1 draw with an alliance and have their victory over you be ‘stronger’ than yours over them leaves a bit of a bitter/confusing taste.
Finally, I will say that the fact that the RR format can allow the existence of 4-robot alliances even in small competitions creates a much less stressful feeling during playoffs, and the requirement to include everyone resulted in all 4 teams being very involved in all the games (even the ones they weren’t participating in) and in the strategic discussions - at a level we never felt in any competition we were in that included 4 team alliances.
I would first of all be happy if more Israeli teams would comment here on their experience and the feeling of the RR format as we held it in the Israeli offseason, to see if they share our feelings, and make this post a place to discuss the best option and have a flood of ideas on how to solve the gaps that exist in each of the formats compared to others. And of course, to hear additional opinions and also from those who may have tried to follow from a distance using the stream (although I know it was not perfect and it is difficult to follow for those who don’t speak Hebrew).
I’m sure there are things I might be missing that could make things a little more difficult with the RR format, but at least for now the advantages we’ve seen seem to greatly outweigh the small drawbacks, so for us the order of preference after experiencing all 3 formats is: RR>standard>DE.
Sorry for the super long post,
Tom – Orbit #1690
