Tower Strength Raised to 10 for CMP

[Doug Frisk]

Originally Posted by JimWright949

I'm hoping Einstein goes up to 11.

Thank you Nigel Tufnel.

[ASmith1675]

Its an interesting rule change, and I think a good one overall.

This might be a slight boon to some quicker low goal bots, but overall what it incentivizes efficiency and teamwork. (I'd be interested to see statistics about how many teams could score 8 boulders solo vs 10 boulders solo). Low goal bots have the advantage of a "miss" usually being quick to recover from. Accurate high goalers still have a decided advantage compared to a low goal bot, but misses are MUCH more costly now.

This will also be a benefit to teams that are willing and able to modify their strategy on the fly to take the relative "sure-thing" low goal while being defended, compared to missing one or more high goals.

[Procolsaurus]

Originally Posted by Caleb Sykes

There are 142 robots attending championships who can consistently score 4 or more boulders per match.* With the tower strength at 8, that would mean that there would probably be enough pairings for all 64 playoff alliances to reliably get enough boulders for captures with just two robots, freeing up the third to do something else.

I had not taken a look at this data until now. I really like it, thank you for providing it.

I just graphed the teleoperated high boulder score, the stat I think best reflects playoff performance. Of course this data cannot be directly applied to champs since teams will behave differently at champs than the events the data is from.

Average calculated teleoperated high boulder score

• 0-1 boulders: 55% of teams
• 1-2 boulders: 12% of teams
• 2-3 boulders: 11% of teams
• 3-4 boulders: 16% of teams
• 4-5 boulders: 6% of teams
• 5 or more boulders: 4% of teams

[Mitchell1714]

Originally Posted by Procolsaurus

I had not taken a look at this data until now. I really like it, thank you for providing it.

I just graphed the teleoperated high boulder score, the stat I think best reflects playoff performance. Of course this data cannot be directly applied to champs since teams will behave differently at champs than the events the data is from.

Average calculated teleoperated high boulder score

• 0-1 boulders: 55% of teams
• 1-2 boulders: 12% of teams
• 2-3 boulders: 11% of teams
• 3-4 boulders: 16% of teams
• 4-5 boulders: 6% of teams
• 5 or more boulders: 4% of teams

The part about this graph that most intrigues me is the single team* that averages 7.5 balls high a match. That is a whole boulder a match more than the second best shooting team.

*I'm going to guess it's 254.

[JamesBrown]

Originally Posted by Mitchell1714

*I'm going to guess it's 254.

We have a winner.

[Richard Wallace]

Originally Posted by Richard Wallace

Einstein will be like the OK Corral at high noon. Better bring a pair of 6 shooters if you expect to survive.

After thinking a bit longer* about Caleb's database, I sorted the teams that have demonstrated 5+ high goals per match, using eOPR1 to rank. I think it will very interesting to see how these teams are distributed across CMP divisions. Alliances with two of these high shooters will be challenging to play against.

The list of 24 25 top guns includes

• nine Michigan teams: 27, 33, 67, 85, 2771, 3620, 3641, 4967, 5460
• five California teams: 254, 330, 971, 3309, 3476
• three Canadian teams: 1241, 2056, 4334
• two three New England teams: 195, 1519, 4564
• and one team each from Indiana (1024), Texas (118), Iowa Minnesota (5172), Arizona Nevada (987), and PNW (1540)

--------

*but not long enough!

[Nathan Streeter]

Originally Posted by Richard Wallace

After thinking a bit longer about Caleb's database, I sorted the teams that have demonstrated 5+ high goals per match, using eOPR1 to rank. I think it will very interesting to see how these teams are distributed across CMP divisions. Alliances with two of these high shooters will be challenging to play against.

The list of 24 top guns includes

• nine Michigan teams: 27, 33, 67, 85, 2771, 3620, 3641, 4967, 5460
• five California teams: 254, 330, 971, 3309, 3476
• three Canadian teams: 1241, 2056, 4334
• two New England teams: 195, 4564
• and one team each from Indiana (1024), Texas (118), Iowa (5172), Arizona (987), and PNW (1540)

1519 also meets this criteria with our performance at New England CMP... our overall OPR was higher at Pine Tree (although our teleop High Goals weren't), so that event's stats are what is included in the CMP Preview tab.

1519 @ NE CMP:

0.73 Auto HG
5.72 Tele HG
14.56 Auto Points
81.79 eOPR1

1519 @ Pine Tree:

0.76 Auto HG
4.65 Tele HG
19.19 Auto Points
82.79 eOPR1

[bstew]

Originally Posted by Richard Wallace

• and one team each from Indiana (1024), Texas (118), Iowa (5172), Arizona (987), and PNW (1540)

While 5172 won the Iowa Regional, they are still from Minnesota.

[Greg Needel]

Originally Posted by Richard Wallace

After thinking a bit longer about Caleb's database, I sorted the teams that have demonstrated 5+ high goals per match, using eOPR1 to rank. I think it will very interesting to see how these teams are distributed across CMP divisions. Alliances with two of these high shooters will be challenging to play against.

The list of 24 top guns includes

• nine Michigan teams: 27, 33, 67, 85, 2771, 3620, 3641, 4967, 5460
• five California teams: 254, 330, 971, 3309, 3476
• three Canadian teams: 1241, 2056, 4334
• two New England teams: 195, 4564
• and one team each from Indiana (1024), Texas (118), Iowa (5172), Arizona (987), and PNW (1540)

This is an interesting data set, but looking at it I think it can't be accepted right at face value.

IMO the 5 balls per match stat is very situational. The two biggest cases of tons of high goal scoring are doing low bar cycles and stealing from the secret passage. In quals for example I think that stealing from the secret passage is going to be much more prevalent than in elims, just because of the defender who will be in play. The low bar scorers who hit 5+ per match are great but how many of them do less if this route is not available (ie. 2 low bar cyclers on a single alliance).

I think that this stat does give us a good top list of the best shooters at the championship, but I also believe we will see some teams on that list fall to defense with lower averages and some teams with lower per shot averages jump up, as the situation changes.



Also in general I feel that the biggest impact of increasing the tower strength at championships will be to combat defense during qualification rounds. There will be less 2 robot alliances who can drop the tower themselves so in order to secure the 4RP mach, the 3rd partner can't be just a defender. On the flip side it also has the possibility to make defense more impactful as keeping an alliance from scoring 1-2 balls can really sway things in regards to event ranking.

[TDav540]

Originally Posted by bstew

While 5172 won the Iowa Regional, they are still from Minnesota.

Originally Posted by Richard Wallace

[*]and one team each from Indiana (1024), Texas (118), Iowa (5172), Arizona (987), and PNW (1540)

Same with 987. Won the Arizona West regional, but are from Las Vegas.

[BethMo]

Originally Posted by Amit3339

the real question that still keeps the capture precent pretty low is will the tower will be captured? Will all the teams will be on the batter? That's the biggest misery

Not sure if this was deliberate or a typo, but it made me giggle. Yes, the biggest misery is when one of your partners doesn't make it onto the batter!

[CJ_Elliott]

Originally Posted by bstew

While 5172 won the Iowa Regional, they are still from Minnesota.

5172 is a force to be reckoned with. anyone who looks over them is a fool

[jajabinx124]

Originally Posted by CJ_Elliott

5172 is a force to be reckoned with. anyone who looks over them is a fool

While I do agree with that, I'm sure they'll make big enough splashes at champs to be seen by good scouting teams or be the alliance captain themselves. I'm excited to see how 5172 performs. Certainly one of the best MN bots going to championships in a while. The last standout one had to be 2169 in 2013 heading into champs.

[MamaSpoldi]

Originally Posted by Richard Wallace

After thinking a bit longer* about Caleb's database, I sorted the teams that have demonstrated 5+ high goals per match, using eOPR1 to rank. I think it will very interesting to see how these teams are distributed across CMP divisions. Alliances with two of these high shooters will be challenging to play against.

The list of 24 25 top guns includes

• nine Michigan teams: 27, 33, 67, 85, 2771, 3620, 3641, 4967, 5460
• five California teams: 254, 330, 971, 3309, 3476
• three Canadian teams: 1241, 2056, 4334
• two three New England teams: 195, 1519, 4564
• and one team each from Indiana (1024), Texas (118), Iowa Minnesota (5172), Arizona Nevada (987), and PNW (1540)

--------

*but not long enough!

Originally Posted by Nathan Streeter

1519 also meets this criteria with our performance at New England CMP... our overall OPR was higher at Pine Tree (although our teleop High Goals weren't), so that event's stats are what is included in the CMP Preview tab.

1519 @ NE CMP:

0.73 Auto HG
5.72 Tele HG
14.56 Auto Points
81.79 eOPR1

1519 @ Pine Tree:

0.76 Auto HG
4.65 Tele HG
19.19 Auto Points
82.79 eOPR1

Not entirely certain of your criteria... or of how these various values were calculated for the table but I believe that Team 230 also qualifies especially if you take into account the "Subtracted Tower Strength" values which are both above 5. Since we haven't been shooting any low goals (although we are able to if needed) this value indicates that we are averaging 5 high goals per match in both of our last 2 events:

230 @ NEDCMP:

0.69 Auto HG
4.99 Tele HG
5.52 Subtracted Tower Strength
16.43 Auto Points
82.86 eOPR1

230 @ Pinetree:

0.66 Auto HG
5.54 Tele HG
5.94 Subtracted Tower Strength
12.25 Auto Points
78.85 eOPR1

(Note our autopoints were lower at Pinetree as we were doing our tuning of our autonomous shot.)

[TechWaffle]

They might have changed it so that it provides more challenge to down the tower and possibly go into the negatives. I like to think this is a good change for Worlds.