http://sanfrancisco.giants.mlb.com/stats/individual_player_gamebygamelog.jsp?c_id=sf&playerID=11 1188&statType=1

There is a definite trend on the supporting staff.  

I guess a stat to find is how many games did our opening day lineup play. See what their average is runs scored per game.

EQA sets OBP at 1.5 times the value of SLG. For the current ERA, OBP is worth around 1.8 times SLG or thereabouts, so EQA "underrates" high OBP players, compared to other linear weights runs estimators that are "tweaked" for the modern era.

"The Kids" is defined as anyone not of the KRAD FM squad

• Bonds did not start one game with three or more of The Kids in the lineup
  
• In the 46 Games that Bonds did not start, only ten games consisted of four or more Vets (22%)
  
• In 36 Bondsless games four or more of The Kids started.

On a side note, this is really disturbing as an evaluation of Bochy.

I'm basically saying that people are making this too complex, trying to tease out the Bonds runs differential issue.

Just forget Bonds, and evaluate the current roster as is.

I'm a longtime lurker--even back to the days at Waiting for Boof when Grant was half-jokingly bleging about being willing to write ufc programs (although I'm pretty sure back then the neologism "bleg" wasn't invented yet). So you haven't noticed my posts before since they haven't existed that long.

I wasn't trying to be dismissive, but rather avert pedantry. Since you are (sort of) passive-agressively attacking my character, however, I feel the need to respond.

"What selective stats did I use?"; "What stats did I cherry pick? The only stat I mentioned in this entire thread was the average OPS of "the kids" vs. "the vets"."

To recapitulate: a supposition was made. It was wrong. I said so. You took issue with that and brought up +- stats. It is a poor stat, used in basketball because of the dearth of stats available. The entire ontology of +- is that it germinated out of a desire to find the best combination of players. The only way Bonds wasn't one of the best combination of players is either: 1) he is much worse than is commonly supposed (and nobody believes that); or 2) he exerts some nebulous force to make his teammates worse when he is on the field--which is ridiculous. I assumed you, like many sports pundits, believed 1). If I was mistaken, I'm sorry. If so, though, there is no reason to even broach +/-. Not that there is one otherwise.

You also repeatedly asked if the Bonds/non-Bonds runs scored per game stat was an indication that possibly the Giants weren't as bad as is often thought. Not to be a jerk, but: no. You are wrong. There have been studies done, and the best way to measure "true" talent is to take into account EVERY data point, not just a few. I pointed to the Cain run support numbers, which every Giant fan should have a rough idea of anyway. You should know the sample size issues with runs scored. In case I need to make the point explicit, here it is:

The Giants scored 101 runs in the 32 games Cain started. But 32 is less than 47. So, in order to make a valid comparison with the non-Bonds games, let's add another 15 games, assuming that the Giants score (their rough average of) four runs per game. Thus, our group of 47 is complete. The number of runs scored is 161, for a per game average of 3.43. This is almost the exact same difference with the average Giants game as the average non-Bonds game is with the average Bonds game. Since Cain has an essentially nil impact on offensive runs scored per game, this should tell you that the games without Barry are likely to be similarly lacking in saying something meaningful about the "true" offensive talent on the field at the time.

How can we be sure? By doing what we should have been doing from the start: examining Barry's actual contributions to Giants wins. There is a stat designed explicitly for this very purpose: Win Shares. According to THT, Barry had 21 win shares last year, effectively giving him credit for 7 wins. But a better method is to measure Win Shares Above Bench; this measures how many win shares Barry would have above a typical bench player. Barry had 12, good for 4 wins above replacement level. Also, both WS and WSAB give credit for defense. Some people think too little, so if you are pessimistic about Barry's D, you could knock off a win. That is an example of an honest analysis of Barry's contributions, not a cherry-picked stat.

Why is the run differential stat dishonest? Because it isn't the best stat available. If you have data for a whole season, you HAVE to look at it. No exceptions. You measure the talent of Bonds (and each of his teammates) by looking at his (their) individual contributions. The only reason people are pointing to this stat (and not, say, the equally meaningless Cain stat) is because it makes the Giants look good. Not honest. Use Win Shares, Linear Weights, etc.--hell, use obp or ba--but use all the data, not only a handful. Hope is not a good method of prediction.

"When did I say Winn/Roberts/Rowand is better than Winn/Roberts/Rowand? "

You didn't; I was giving a hypothetical example of what one would have to do to make an honest analysis that the Giants will be better than last year (and something very similar would have to be done to "show" the Giants are likely to be better than most assume). I disagree with your analysis, but what you did in the last post is what I was saying should be attempted.

Goofus, I like you as a poster and you are a respected member of the community, so I am not trying to pick on you. "Statheads" were asked to answer a question and I responded. I'm sorry if I misrepresented your posts, but if I did so I legitimately didn't understand what you were trying to say. We all hope that the Giants will be better next year, and I don't fault you for that. But the stats that you pointed to base your (potential) optomism on do not support your conclusions, however tentative.