Login

icewindoo · 09-15-2024, 11:07 PM

My player is a right-handed D who played this past season on the left side of the D pairing, and this got me to wondering: is there a drawback to D playing on their "off" side compared to their "natural" side? For example, is it better for right-handed D to play in the RD position, or does it not really matter?

Answering this question is hard first and foremost because it's hard to even tell the difference between "good" and "bad" performance by players. Lucky for me, I've started using the Player Contribution (PC) system to analyze players, and despite its many flaws, I consider it to be a useful way to reduce the quality of a player's overall on-ice performance down to a single number.

Details on the PC system and the player cards I create can be found in another post here, but here's a sample player card from this past season:

[Image: X9FyvJn.png]

To answer the question at hand, I looked at the overall PC values for all SHL D over the past eight seasons. I split the data into six categories:

all D who played in the LD position
all D who played in the RD position
left-handed D who played in the LD position
right-handed D who played in the LD position
left-handed D who played in the RD position
right-handed D who played in the RD position

The PC values on their own are useful, but I made an additional correction to clean up the data. The number of data points here is not huge: only 120 D play per season. It's certainly possible that there's some bias in position selection, for example, like high-TPE players being more likely to play in their natural position. I don't know if that's true, but I tried to correct for any similar bias by calculating a normalized value of PC divided by attribute quality. That is, take the quality of the performance shown by the stats (PC), and divide it by how many attribute points they have. I didn't use TPE directly, as I've seen that "total attribute points" is somewhat more predictive of player quality.

Anyway, enough ham-handed words about numbers. Here's the picture that matters:

[Image: sKMQT3L.png]

This chart suggests to me that there is indeed a negative impact of playing on the "off" side, and this effect is consistently repeated from season to season, and for both LD and RD separately, though to varying degrees. For example, in S75, left-handed D playing LD and right-handed D playing RD far outperformed their comrades who were playing on the "off" side. There are a few cases where the pattern doesn't hold completely; for example, in S74 right-handed D playing LD had higher PC values than the right-handed D playing RD, but only by a very small amount. But in almost every scenario, the pattern holds: switching to the "off" side seems to reduce the player's contribution. The average amount of reduction measured here is about 9%.

Not noted in the graph is that about 61% of the D in this data set are right-handed. That's not a huge surprise to me, since that works out to be roughly half-way between real-life hockey handedness and real-life writing handedness. But what is a surprise to me is that only 50% of left-handed D are played on their natural side (LD) and only 51% of right-handed D are played on their natural side (RD); it seems clear that on the whole, SHL GMs completely ignore the defender's handedness when choosing positions.

Anyway, my own conclusion that this analysis does indeed hold a meaningful insight into player performance depends highly on my reliance on the PC system. And obviously, I'm not a statistician, so I can't really say whether enough data is used here to make the variation in results statistically significant. But I thought it an interesting result, and am happy to share.

Bongo · 09-16-2024, 09:54 PM

excellent analysis..

***Anthique*** · 09-17-2024, 08:52 PM

Gonna have to try this with Team Canada now Smile

the5urreal · 09-17-2024, 09:01 PM

Love this k8nd of data analysis. Now do wingers!

Navigation

Extra Menu

About us