In short: let's use the ODL example. I set some baseline minimums for the aggregate percentages. I want to end up with a minutes adjusted team rebounding number of >= 90/30, assists at >= 60, team eFG% of >= 54, defense >= 70, turnover <=14 (plus minimizing fouls as much as I can). Since we only draft 7 players, I need to end up with numbers that are above that since my bench will drag it down. I typically want to end the draft with around 14000 - 15000 minutes and $150k/min/82 left to round out the bench.

I am usually trying to get one guard and one forward/center in the first two rounds. By the end of round 3, I usually want at least 8 points of usage. I hopefully have one paint monster and one perimeter monster at that point as well. By the end of round five, I want three 80+ defenders capable of covering all five positions in terms of effectiveness, but sometimes I will dip on the third defender.

As the first couple rounds, I look at who is available and figure out which of the 1-2 stats from above I will excel at. I will try and pump that up as much as I can and use the final rounds to make sure I cover up the weaknesses. I like to try and get my bench usage covered by the end of the draft as well.

For larger caps, it is the same except the baseline is higher and I will typically end up with way more minutes drafted knowing I can spend more in the draft and won't need to use as many castoff minutes with the undrafted guys.

There are some obvious things I do in terms in lineup optimization and stuff but that is the general draft approach. I have built a lot of automated tools that suggest players to me as the draft unfolds and stuff which is helpful. They also keep a running total of the finances and show me the likely bench players I will have to round out the roster. But that is mostly just time saving stuff.

I don't really have a great system for uncapped leagues at this point in time.

Posted by raggedclaws on 12/22/2020 1:46:00 PM (view original):
I am curious about how shot location % is accounted for in price, if at all. 94-95 (and other) Rodmans, 66-67 (and other) Wilts, most Dwight years - these seasons are all very good for their own reasons, but I find them especially valuable because none of them shoot more than 15% from the midrange. There are equivalents from the perimeter, players like 15-16 or 18-19 Curry, most Korvers, Duncan Robinson, some Jingles, some Lowrys.

Obviously it's no secret that there's value in varying shot selection, but a team consisting of a combination of interior and perimeter players like I listed above is clearly better than a mid-range focused team with the same stats, since many teams never adjust their defensive positioning, and if you have enough variance it can be even worse for your opponent to pick a direction to shift in, leaving shooters or dunkers more open, not to mention the increased foul%.

I wonder if that PNT/PER value is incorporated into player salaries or if there's some Moneyball potential there as well

When I crunched the data, it looked like it was in there but weighted pretty lightly compared to everything else.