NBA formula

[SiteWolf]

So, we all know the NBA formulas have sucked since they sucked even more in the very beginning. Those original formulas even helped guys like Relic figure some things out about our formulas I'm not sure they would have otherwise....but it is what it is, that water has flowed already.

Here's the problem with the NBA here

In EVERY other sport we track, there's stat categories where a player either did or didn't do something. A WR or RB either did or didn't score a touchdown...yards may only lessen the negative or create a smaller positive. Similar in baseball- a hitter either did or didn't hit a home run or steal a base. Hockey- goal or no goal. Golf- eagle or no eagle. NASCAR- top 5/10 or no. Other ancillary stats help of course, but at lower levels.

None of that is a news flash to any of you. Certain stat categories effect dividends more than others. My whole point with formulas is to try and somewhat emulate a players value in a game with their dividend- at least as much as can happen using statistics that appear in box scores.

Basketball is simply different. If a player plays, they'll generally generate stats in almost ALL the categories we track. Splitting out by position doesn't help us. A guard will still get rebounds, a center will still get assists, their projections will determine things more than different formulas by position would.

While every other sport has categories that can cause more volatility in dividends and give us the effect ideal here, giving reason to buy and sell stocks on a specific day, basketball, well, to graph that market you'd see a lot of small hills and very few mountains and canyons.

Answer? Well, it would be possible to input divi bumps when a player gets 30 points, 15 rebounds, that sort of thing. It would be possible to create even more elaborate formulas. But I'm not sure how much that actually helps. Plus, it's not something I can do without Nick's help and his time has been severely stretched.

There are tweaks that I can and do do, but one I've been thinking about recently is.....what if we just narrowed the minimums and maximums of projections? Raise the floor, but drop the ceiling? A player with low projections with a tendency to go off for a game here and there would yield more negatives with the occasional bigger dividend, while the 'blue chipper' would at least have a chance to generate a nice divi once in awhile, instead of a 40 point game barely being a buy........I think.

[Yami Ron]

How about this Orren:

A player must play in 70% of a season's games to qualify for league leader stat. Also a player must make at least 300 field goals, 125 free throws, and 82 threes to qualify for percentages.

So in regards to the projections, what you can do from season to season is make all the projections for a player 70%-75% of what they did last year.

I think another negative stat might be needed: Missed Free Throws

Then for game minimums based on the 70% above or for rookies:

8 Points, 1.5 Threes, 4 Rebounds, 4 Assists, 1.25 Steals, 1.25 Blocks, 1 Turnover, 1 Missed Free Throw across the board, regardless of position.

Multiplying that by 82:

656 Points, 123 Threes, 328 Rebounds, 328 Assists, 103 Steals, 103 Blocks, 82 Turnovers, 82 Missed Free Throws (projection only counts if they attempted a FT in that game)

Use a player's lower 70% stat values in Threes, Assists, Steals, Blocks, Turnovers, and Missed Free Throws. (If any 70% stat value falls below 1, use 1 in those situations)

You can add a hidden divi percentage bonus if the player gets a double or triple double. I don't think you need to add a bonus for a specific stat.

[SiteWolf]

I appreciate the thoughts.
Some years I don't, but so many years I've spent time playing with new formulas and thinking about creative ways to do things...and nothing i've come up with improves things enough to blow up what we've already had. And, unfortunately, for the moment I can only plan on changes I can do without programming.

We soldier on!

[SiteWolf]

Not that I know what to do with this information, but I compiled some numbers from last season and created average numbers in 3 tiers.

	points	threepointers	rebounds	assists	steals	blocks
Tier 1	22	2	7	5	1	1
	1822	161	573	414	86	49
						
Tier 2	13	2	5	3	1	1
	1086	133	415	241	73	48
						
Tier 3	9	1	5	2	1	1
	771	80	373	159	55	48

[rich76]

The categories that stick out are 3's and blocks. Not even sure how to understand that.

[SiteWolf]

Originally Posted by rich76

The categories that stick out are 3's and blocks. Not even sure how to understand that.

oh definitely....IF I were to really do something with stats like that, I'd probably need to do it on a per category basis....this was just a quick thing where I totaled points thru steals and sorted those combined numbers into these tiers- top 50 in tier 1, next 100 in 2, rest with solid number of games played in 3