2018-19 team win projection contest

Home for all your discussion of basketball statistical analysis.
Crow
Posts: 5387
Joined: Thu Apr 14, 2011 11:10 pm

2018-19 team win projection contest

Post by Crow » Sat Aug 04, 2018 9:50 pm

Some early projections have been released.

https://mobile.twitter.com/JacobEGoldst ... 17/photo/1


You can put yours here if / when you want (update as you wish).

Commentary on the topic is fine too.

eminence
Posts: 64
Joined: Sun Sep 10, 2017 8:20 pm

Re: 2018-19 team win projection contest

Post by eminence » Sun Aug 05, 2018 11:04 am

Let's make it a repeat, woohoo!

Teams I'm particularly interested in seeing where people put: Pacers, Pelicans, Nets

Pelton seems to have a very tight spread on his RPM projections

Crow
Posts: 5387
Joined: Thu Apr 14, 2011 11:10 pm

Re: 2018-19 team win projection contest

Post by Crow » Sun Aug 05, 2018 3:25 pm

Yes, Pelton's numbers are pretty tight and I'd favor a broader distribution.

I am pretty favorable to Pacers but haven't picked a win total yet. I'd go low or real low on Nets. Pelicans, I dunno. Probably middle of road.

eminence
Posts: 64
Joined: Sun Sep 10, 2017 8:20 pm

Re: 2018-19 team win projection contest

Post by eminence » Sun Aug 05, 2018 4:11 pm

I'd say I'm fairly high on both Pacers/Pelicans (bit over 50), contemplating both for home court. Nets I seem reasonably lower on than most, they'll finally have their own pick back right?

Crow
Posts: 5387
Joined: Thu Apr 14, 2011 11:10 pm

Re: 2018-19 team win projection contest

Post by Crow » Sun Aug 05, 2018 5:20 pm

Yes.

Crow
Posts: 5387
Joined: Thu Apr 14, 2011 11:10 pm

Re: 2018-19 team win projection contest

Post by Crow » Sun Aug 05, 2018 7:40 pm

Back in April you said it might be wise to pull extreme highs down a little and weight late season performance more heavily. Still intend to implement these changes? Anything about research supporting these thoughts you want to share? Were these statements purely based on last season or multi-year? Re: late season performance, how many performance clusters are there and what are the trends for teams based on early season performance or other criteria?

On average the projections were almost 2 wins high in west, a little over 1 low in east last season (if I used Mike's data right). For actual wins, the east improved by almost 1.5 wins in 17-18 over 16-17 and the west declined by similar amount. It appears the predictions missed this trend or noise. What divisional trend do people expect for next season? Does anyone use a standalone coaching adjustment beyond the implicit coaching impact in player performance?

eminence
Posts: 64
Joined: Sun Sep 10, 2017 8:20 pm

Re: 2018-19 team win projection contest

Post by eminence » Sun Aug 05, 2018 8:09 pm

Crow wrote:
Sun Aug 05, 2018 7:40 pm
Back in April you said it might be wise to pull extreme highs down a little and weight late season performance more heavily. Still intend to implement these changes? Anything about research supporting these thoughts you want to share? Were these statements purely based on last season or multi-year? Re: late season performance, how many performance clusters are there and what are the trends for teams based on early season performance or other criteria?

On average the projections were almost 2 wins high in west, a little over 1 low in east last season (if I used Mike's data right). For actual wins, the east improved by almost 1.5 wins in 17-18 over 16-17 and the west declined by similar amount. It appears the predictions missed this trend or noise. What divisional trend do people expect for next season?
At the time they were just gut feelings, but after looking at it a bit more I'll stand by pulling the extremely high win projections down a bit (I had GS at 70 wins last year, lol), I didn't keep the numbers, but it really only seemed to apply at 65+ wins. I don't think I actually have anybody quite that high this season at first pass.

The late season thing is still gut, haven't looked at it any more. I have a completely subjective directional variable in my projection formula and I'll just consider it there I think.

One thing I have introduced in my projections is making them bimodal, with playoff teams pulled towards ~50 wins and lottery squads pulled towards ~30.

I'd imagine the trend largely reverses next season, bottom of the West really seems to have taken a step forward this season.

Crow
Posts: 5387
Joined: Thu Apr 14, 2011 11:10 pm

Re: 2018-19 team win projection contest

Post by Crow » Sun Aug 05, 2018 8:38 pm

Thanks for replies.

Bottom of west probably has improved compared to last season and relative to bottom of east but my quick take is that top of east may improved relative to last season and top of west. The net effect probably isn't large but haven't looked close enough to guess at net overall tilt. May not need to but it could be revelant for finalization / rounding.

Cicada
Posts: 5
Joined: Mon Aug 06, 2018 2:59 am

Re: 2018-19 team win projection contest

Post by Cicada » Tue Aug 07, 2018 12:49 am

Anyone have any minutes projections they’d be willing to share? I’ve seen 538 has their minutes projections available but am still looking for more resources to compare to.

Crow
Posts: 5387
Joined: Thu Apr 14, 2011 11:10 pm

Re: 2018-19 team win projection contest

Post by Crow » Tue Aug 07, 2018 1:37 am

Kevin Pelton has shared his here in past. Did so early last October. Perhaps he'd do so earlier if asked (rec, do so on twitter).

Dan Dickey (hoops underscore nerd on twitter), Josh Engleman, daily fantasy sites... might find something.

Basketball-reference will when they release their player projections (using weighted 3 year average). Probably not summarized though.

The upcoming BasketballIndex.com might share with subscribers.

DSMok1
Posts: 848
Joined: Thu Apr 14, 2011 11:18 pm
Location: Maine
Contact:

Re: 2018-19 team win projection contest

Post by DSMok1 » Tue Aug 07, 2018 5:23 pm

Here are Jacob Goldstein's live projections:



https://docs.google.com/forms/d/e/1FAIp ... g/viewform
Developer of Box Plus/Minus
APBRmetrics Forum Administrator
GodismyJudgeOK.com/DStats/
Twitter.com/DSMok1

shadow
Posts: 200
Joined: Fri May 29, 2015 3:38 am

Re: 2018-19 team win projection contest

Post by shadow » Wed Aug 08, 2018 5:59 pm

Vegas Westgate win totals via twitter:

Code: Select all

Atl 23.5
Bos 57.5
Brk 32.5
Cha 35.5
Chi 28.5
Cle 30.5
Dal 34.5
Den 47.5
Det 37.5
GS  62.5
Hou 54.5
Ind 47.5
LAC 35.5
LAL 48
Mem 34.5
Mia 41.5
Mil 46.5
Min 44.5
NO  45.5
NY  29.5
OKC 50.5
Orl 31.5
Phi 54.5
Phx 29.5
Por 41.5
Sac 25.5
SA  44
Tor 54.5
Uth 48.5
Wsh 44.5
Last edited by DSMok1 on Wed Aug 08, 2018 7:39 pm, edited 1 time in total.
Reason: Fixed formatting


Crow
Posts: 5387
Joined: Thu Apr 14, 2011 11:10 pm

Re: 2018-19 team win projection contest

Post by Crow » Thu Aug 23, 2018 1:58 am

I agree more with 538 on the Mavs than the other early projections which are higher.

jgoldstein34
Posts: 240
Joined: Fri Sep 30, 2016 6:38 pm

Re: 2018-19 team win projection contest

Post by jgoldstein34 » Tue Aug 28, 2018 9:44 pm

Seems the biggest difference between my projections and 538 for the Mavs is how they value Doncic at around replacement level (-2.2) while my model has him as a small positive (+0.4)

Post Reply