R$ and 5x5 values missing

[Kevin]

Noticed that there are no dollar values being generated by the Stats and Projections section... Glitch?

Sample:

http://www.baseballhq.com/members/tools/projections/proj.php?lg=A&p=B&t=PY

[RobR]

It's the time of the year where we run out of players with a significant playing time remaining to fill out a full fantasy squad. It's similar to the situation at the beginning of the year where we need enough players to get playing time before we can generate YTD dollar values.

For the dollar values in our projections files and PlayerLink we require that there be 9 pitchers per team (11 AL, 12 NL) that are projected for 6 or more IP. We also require that there be enough batters per team (2 C, 1 1B, 1 2B, 1 SS, 1 3B, 1 CI, 1 MI, 5 OF, 1 UT) projected for 21 or more AB.

[Kevin]

Why not assume the base rates and replacement-level values from the league-wide YTD values?

[RobR]

Assuming there was value in calculating projected dollar values at this point in the season, I think there are possible approaches to consider for next season. But none of them are coded right now. So I'm sorry, but I can't offer anything other than an explanation now.

I think small sample size valuations are very noisy (e.g, the valuation difference between 7 and 8 projected RBI at this point) and lose value as the noise increases with decreasing sample size. Players' value fluctuates significantly depending on each fantasy team and their adjacent compeitors' standing in each category. While there could be some value in dollar values at this point, I think owners are better off evaluating players' projected stats in the context of their league standings.

[Kevin]

While there could be some value in dollar values at this point, I think owners are better off evaluating players' projected stats in the context of their league standings.

I agree with you that much of the time that is better, but do you agree with me that there might be at least some non-zero merit in looking at projected dollar values? It seems we would be better off having many different tools at our disposal, instead of having one type of information removed because of how HQ would prefer us to manage our teams.

Perhaps I'm in a bit of a rush Sunday night and I'd like to start my FAAB process by at least sorting the players by projected 2007 value, before zeroing in on the characters with the right mix of categories (by the way, this is a true story, based in part on my own life). Or, I'm totally out of the race and while I don't care about the current standings I have some roster spots I'd like to use to take a shot on keepers with the best combinations of BPVs and playing time (often well approximated by $$$). In either case, it sure would save me a ton of time to be able to sort by $$$ as I begin my process. I just don't have as much time to do this stuff as I did in college I'm afraid; your typical HQ member may be different (maybe many of them are retired? ).

Although I appreciate your advice to look at my league standings in September, and it is good advice which I plan to follow, please take me at my word that saving me some time in terms of filtering and sorting would be as big a help in September as it is earlier in the year. If you can find it palatable to provide some sort of avenue to this information next year and beyond, it would be a helpful complement to the above advice. Thanks man.

[RobR]

No argument with you that more information is better than less and I certainly agree there's non-zero merit in what you suggest. At the same time I'd feel bad if I didn't warn folks about the small sample size dollar values. With resources limited, while our cutoff for tasks to consider might be those with non-zero merit, I couldn't guess right now whether this item will be high enough up Ron's list for next year.

For next year I was planning on (pending Ron's approval) adding Projected Total $ values to Playerlink. Perhaps this could work its way into the Proj + YTD projections and stats page, too? Even better than the projected dollar values for the remainder of the season might be the difference between Proj + YTD $ and YTD $ (and vice versa at the start of the season). That difference might yield better information of the same type you're looking for. What do you think?

[Howie M]

In a way, what people REALLY want is whether the $6 or $20 player all year is likely to continue that pace in September.

Really meaning - playing time guesses.
I'm in an NL league where I can nearly nail down all the info on who might play a lot in Sept.
And we have limited 'free' moves anyway.

But the other extreme - mixed league, daily moves, high stakes, rich but very busy HQ subscriber - well, he'd KILL to find out that "Joe Blow" will get an extended look at the expense of that owner's worst regular/player.

There might be a compromise there somewhere.

[Kevin]

Even better than the projected dollar values for the remainder of the season might be the difference between Proj + YTD $ and YTD $ (and vice versa at the start of the season). That difference might yield better information of the same type you're looking for. What do you think?

I would be fine with anything that allows me to quickly sort and filter by expected contribution over the rest of the year... It sounds like what you propose sort of comes close, and I could probably engineer a solution in Excel to back out the Proj figures. Maybe that would be enough extra steps to take such that HQ won't feel responsible for people accidentally setting fire to their homes with the small sample size info.

[RON@HQ]

What I've always done in September is sort all players by projected expected playing time, filter out those that do not meet an AB/IP minimum that I set, and then sort by BPV. Whether we project someone to get 30 AB during September or 40 is pretty much a shot in the dark so I'd rather just have them all in the same pool of potential targets.

[Howie M]

Guh, Shandler is one move ahead of me again!

[Kevin]

How's this for a compromise - just freeze the values as of the last time that HQ is willing to calculate projected $$$s, and make them available in a data column labeled "Aug 31 Proj $" or something similar. Wouldn't require any tweaking of the valuation algorithms/methodologies, yet would still allow members to sort/filter on the HTML page without having to download the .csv files and manipulate the data in Excel to achieve the same effect.

[slaman992003]

Thanks for these posts, I had been wondering why I was always considering guys around 0$ value!! I had resorted to using my Rotolab draft section and the Fantasy point values (which is closer to my league anyway). This takes the stats projected and creates value and ranks them. Somehow it seems the BBHQ site should be able to do something like this. But I'm in first place due to BBHQ, so...it's really hard to complain. But perhaps the Rotolab suggestion may help some of you.

I also like the idea of looking at the BPV of the key guys you are interested in. I do a separate sheet each week and look at the BPV's. Between this and the xAvg, I usually can make a reasonable decision.

[Kevin]

But I'm in first place due to BBHQ, so...it's really hard to complain.

It's not complaining - it's feedback and constructive criticism!