The "rectangular measure", a territory size statistic

[bugcat]

Well then, let’s try correlating rectangular measure with game result.

We’ll use three different measures.

1. Regular rectangular measure (RM)

2. Modified rectangular measure (MRM). MRM = RM x G, in which G is the number of living groups that player has.

The games we’ll examine are ten victories from @seequ (cq), considering only games with margins of less than fifty points.

op	date	mg.	cq’s RM	op’s RM	cq’s MRM	op’s MRM	RM & MRM agreement
StinkyTofu	04-20	6.5	45 (>)	32	135 (>)	96	>
danjosely	04-13	28.5	77 (>)	63	154 (>)	63	>
rocco	04-10	cq	50 (<)	55	150 (>)	110	–
genban	04-08	6.5	30 (<)	36	30 (<)	72	<
Mauled	04-07	11.5	30 (<)	35	90 (>)	70	–
Caeldron	04-04	32.5	40 (>)	21	80 (<)	84	–
KoBa	04-03	26.5	42 (<)	36	42 (<)	108	<
StinkyTofu	03-22	23.5	50 (>)	45	100 (<)	135	–
danjosely	03-18	10.5	40 (<)	38	120 (>)	38	–
marcmarc	03-18	6.5	48 (<)	70	240 (>)	70	–

Success rate:

RM 40%
MRM 60%
RM & MRM agreement 50%

So, thus far we can conclude that RM isn’t a helpful measure for predicting victory.

[claire_yang]

I feel this is not a very good measure for fighting style players, where games would just ended in resign with one side lost the semeai fight.

How do you measure rectangular in these kinds of games?

[bugcat]

I originally meant for RM to be attained from scored positions.

[Atorrante]

@bugcat

RM isn’t a helpful measure for predicting victory

@bugcat

originally meant for RM to be attained from scored positions.

I am getting confused.
How can a measure that is acquired after the game is finished have a predicting value?
Did I miss something?

[bugcat]

People asked me whether RM worked as a predictive measure, so I used RM on some scored positions to assess whether there was a correlation between RM and game result.

Applying to RM to midgame positions is someone else’s job~

[Atorrante]

Okay.
Really appreciate it that you have that careless attitude of “maybe this will result in nothing useful after all, but I am going to do it anyhow (just for the fun of it)”.

[bugcat]

You don’t know until you try, right?

[Vsotvep]

I’m really tempted to play some gote moves to maximise my RM in some of my games now… That one useless move would provide 10 more points!

[shinuito]

I’ll pitch in another data point.

#	Me	opp	W/L	Measure Me	Measure Opp	Game	Score	Komi	SWK
1	2k	7k	W	90=17*4	40=10*4	32374501	B+47.5	6.5	54
2	2k	5k	W	40=10*4	36=9*4	32697469	W+16.5	6.5	10
3	2k	3k	W	48=8*6	48=12*4	32625496	W+12.5	7.5	5
4	2k	2d	L	36=9*4	42=14*3	32621275	W+22.5	7.5	15
5	2k	3d	L	22=11*2	39=13*3	32597116	W+16.5	7.5	9
6	2k	3k	W	50=10*5	56=8*7	32592757	W+8.5	7.5	1
7	2k	?	W	48=12*4	45=9*5	32569245	B+3.5	7.5	11
8	5k	?	W	84=14*6	42=14*3	32564068	W+48.5	7.5	41

SWK=Score without komi.

I’ve added komi to Blacks score and subtracted it from White depending on who won. I think that would reflect more of who was winning on the board, and it’s an extra column which might be of interest.

Game 6 above has the winning player with a smaller rectangle, but actually I counted two non-overlapping rectangles of size 50, so maybe that makes up for it

It’s just a quick ish count, trying the biggest sized rectangles I could spot for both and then moving on. You can double check if you want.

I might add more later if I think of it.

Current average is 52.25. I’m not if the average will mean much really since

@david265

I would imagine that rectangular measure would not correlate with your rank, since you play against players of various ranks and there are all sorts of variations in each game that may not average out in any meaningful way.

and maybe just comparing who has the biggest rectangle vs who won/lost and by how much could be interesting

@Samraku

To echo what has been mentioned before, I’d be interested in how good of a predictor for who has more points who has the larger rectangular measure is.

I probably won’t do the

@bugcat

Modified rectangular measure (MRM). MRM = RM x G, in which G is the number of living groups that player has.

because that seems like hassle. But feel free to copy the table and add in/remove numbers.

Raw table - easier to copy/pase possibly

|#|Me|opp|W/L|Measure Me|Measure Opp|Game|Score|Komi|SWK|
|---|---|---|---|---|---|---|---|---|---|
|1|2k|7k|W|90=```17*4```|40=```10*4```|[32374501](https://online-go.com/game/32374501)|B+47.5|6.5|54|
|2|2k|5k|W|40=```10*4```|36=```9*4```|[32697469](https://online-go.com/game/32697469)|W+16.5|6.5|10|
|3|2k|3k|W|48=```8*6```|48=```12*4```|[32625496](https://online-go.com/game/32625496)|W+12.5|7.5|5|
|4|2k|2d|L|36=```9*4```|42=```14*3```|[32621275](https://online-go.com/game/32621275)|W+22.5|7.5|15|
|5|2k|3d|L|22=```11*2```|39=```13*3```|[32597116](https://online-go.com/game/32597116)|W+16.5|7.5|9|
|6|2k|3k|W|50=```10*5```|56=```8*7```|[32592757](https://online-go.com/game/32592757)|W+8.5|7.5|1|
|7|2k|?|W|48=```12*4```|45=```9*5```|[32569245](https://online-go.com/game/32592757)|B+3.5|7.5|11|
|8|5k|?|W|84=```14*6```|42=```14*3```|[32564068](https://online-go.com/game/32564068)|W+48.5|7.5|41|