I'm looking for some technical help here if anyone has knowledge with this type of thing. I am in a 17 team CBS keeper/auction league. It is a 4x4 league. We have 40 man rosters and we start 15 hitters and 9 pitchers.
Now to the point...based on recent history, I have set goals for each of the 8 categories. For example, I know that 160 SB for the year will put me in the top 3 or 4 in the league in that category. Is there a program or query that I can run in Excel or another software that will give me the optimal line up within my budget constraints?
For example, let's see that I need to auction for 90 more SB, 180 more HR, and 650 more RBI to meet my goals. I also have a 1B, CI, SS, MI, and four OF slots open. If I projected auction prices for those that will be available in the auction, can I run a program that will tell me what my best combination of players is?
Hopefully this makes sense. There is a guy in our league that has a program that generates a line up, but he's not sharing his secrets!
I've never heard of a program that does that, but it sounds fancy. If you can't find a program there a couple of ways you can game plan your auction that will do almost the same thing. One thing you could do is to average those totals over the remaining number of players, and adjust that average as you go. That will give you an 'at a glance' feel for what you need to draft more of. You also need to keep in mind that it's way easier to find speed on the waiver wire, so filling out the power cats first should be a priority. You could also probably do something with the baseball monster values as it relates to cost, but that could get complicated.
As far as 'best combinations' go, you could always try the old fashioned approach and do a bunch of research. Pretty much every site has a list of sleepers and undervalued players you can target, and a quick glance at some competing player projections will give you a good idea where you have some wiggle room in the values. There's so much potential variation in the actual bids on players in any given auction that I think a program would have a tough time quantifying it really, except to do it straight from some list of projected prices. That sounds neat, but once the dollar values start fluctuating, like they do in most auctions, those projections become far less useful. There's really no replacement for solid draft prep the old fashioned way.
Keep in mind that the lineup you draft won't be the lineup you finish with. Certain stats are easier to come by throughout the year, as are different positions. If you know you can find steals and saves on the waiver wire, then lower your pre-draft category goals for those stats and bump up the other ones. All you need to have are proper dollar values and category goals. Finding pockets of value within a draft is much more important than using an algorithm to find the optimal theoretical lineup. Good balance and dollar values will take care of that.
Skin Blues wrote:Keep in mind that the lineup you draft won't be the lineup you finish with. Certain stats are easier to come by throughout the year, as are different positions. If you know you can find steals and saves on the waiver wire, then lower your pre-draft category goals for those stats and bump up the other ones. All you need to have are proper dollar values and category goals. Finding pockets of value within a draft is much more important than using an algorithm to find the optimal theoretical lineup. Good balance and dollar values will take care of that.
I agree...in this league, with 680 players being drafted, there is NOTHING on the wire. It is very tough to find anything than an occasional starter on the wire.
I appreciate your insight though...I think you are definitely right.
Yeah I guess there's not much out there with so many players rostered... in any case, I don't think algorithms for this kind of thing would be as useful as intuition because it wouldn't allow for flexibility. In my opinion it would just be a false sense of security.
I don't think an algorthm can manage a 680 player auction draft either. Much better to rely on your own instinct bolstered, one would hope, by the insticts of a variety of experts. With that many guys drafted you're going to need to be all over the sleepers and late round value picks too.
Fenris-77 wrote:I don't think an algorthm can manage a 680 player auction draft either. Much better to rely on your own instinct bolstered, one would hope, by the insticts of a variety of experts. With that many guys drafted you're going to need to be all over the sleepers and late round value picks too.
Thanks, man. I really do appreciate your thoughts. I think you are right...instincts are the way to go. I did think the algorithm would be an interesting way of looking at it, but when it comes right down to it, it's all about going with your gut.
The Razzball draft tracker is kind of like that. But you wouldn't want to use it in an online snake draft. They have something on the user input page where you input the stats you are trying to acquire, and I believe it's on the war room page where it shows you what your team is averaging per player versus what your input stats says you should be averaging to reach those stats, if that makes sense.
It however won't give you a list of players you want to target, which is what you are looking for, I'm sure there are other draft tools out there, which probably do the same thing as Razzball's tracker.
Excel has the power you need for something like this, at least at a crude level. I've created some nice draft tracking spreadsheets in the past and I know there are some available. None do what you're asking for, but IMO it only takes a little adjustment to do so.
As the draft goes, you check off which players are on your team, and which players have been drafted. I do this by putting a 1 for my team in the draft column, and a specific number for each other team. On another tab, using index matches, I auto populate my team along with their projections. The overall team stats are calculated at the bottom, highlighting if I'm above or below my targets cats and by how much. I also populate each team and have an overall league standings tab that I check as the draft goes on. So I will see how I'm stacking up against other teams during the draft.
I used to have a tab with the top projected remaining players at each position, and highlighting players that are above my target cats. You could have a tab with the list of players above each cat. For example if you want to average 20 Hrs, but you're only at 17, you could quickly see the list of players that are above that rate at each position. Probably best to use filters rather than another tab like I did, but hey we're always learning.
Of course if you draft one of those 20+ HR players, your other stats may be negatively influenced, e.g. SB rate or avg. What is needed is a calculation for all remaining players that would tell you which remaining player would add the most to your overall team rating. So you could have a list of the top players to pick. This last piece is the most complex, but really not that far beyond what I've already done with spreadsheets. If you can calculate overall ranks you can calculate next best available players.
However this would only show who would influence your cats the best, but may not really be the best player to take, especially early. It may be better to take that 30HR guy in the 8th that would lower your avg and Sb rate, and then make up for it later with other players. It would probably be more useful in later rounds to fill out the last starting roster slots, or tell you which pitchers can get you the K rate you need.
So IMO it is doable, at least at a crude level for next best available players. Since it's based on projections that calculation would be only 70% accurate at best, and would not handle the chaos of predicting beyond the next move. Maybe the MIT guys could do a better job.
This would be a bit of work to setup, and I don't use this level of complexity anymore. I found that during drafts just checking off the drafted players was a lot of work. Not so much at the start, but by the 5th or 6th round, someone will draft a player that you have ranked for the 14th round and you have to spend time hunting him down to enter the draft result. By the time you do that another player or two have been taken, and you have to hunt those guys down. Drafts can move along fast. I found that to be a lot of work, and I'd rather enjoy the draft and be able to evaluate what I want to do instead of doing data entry.