Crucial Situations
by Tangotiger
This is going to be one of those articles that starts off slow (boring definitions), has no real meat (why clutter up your reading time with one thousand words when...), but will end up with a bang (... a picture is worth as much). Bear with me, and we'll get there.
What is a clutch (read: pressure, crucial, high-leverage) situation? It is that specific situation where the impact of an event will be much greater than in a typical situation. Man on base, tie game, bottom of the ninth qualifies. Bases empty, down by 8, two outs does not.
A little while ago, I introduced a Win Expectancy chart. This chart showed the chances of a team winning in various score/inning/base/out situations. This is nothing new, as the Mills Brothers generated something similar in the 70s, and many authors have followed suit. It all boils down to the same thing. Winning.
In a continuing look at using Win Expectancy, I will identify the crucial game situations. These situations are for 1 plate appearance only. This is a similar concept as to when to walk Barry Bonds, at home and on the road. They are useful to determine when to bring in a good pinch-hitter. They are not useful to determine when to bring in a good reliever, unless you are planning to only bring in a reliever for one or two batters. I'm currently working on identifying the optimal point in the game (given inning/score/base/out situation) to bring in your best reliever to the end of the inning, or to the end of the following inning.
As you bring up the chart, you will see that many crucial situations occur early in the game. This does not mean that you need to start switching pitchers or batters at that point. But it does mean that the strategies you do make will have much greater impact than in other game situations.
Enjoy the Crucial Game Situations Chart.
December 3, 2002 - David Smyth
Cool charts, Tango. I'm not getting the color in your description. Any idea on how to get the red and grey, etc.?
December 3, 2002 - tangotiger
(www)
(e-mail)
Really? Hmmm, are you using an old version of Netscape? What's your browser version?
December 3, 2002 - jimd
Charts are always cool (when well done). No problem seeing them, but the shading doesn't print for me. Just pages of empty grids. Oh well.
Tango, how about other run situations? In some innings, does the blue shading bleed into the -4/+4 run columns (or even further)?
December 3, 2002 - Doug
Well, I can see the chart and the colors just fine, although the column headings would be better if repeated every half-inning, not just every inning.
But is there anything here that isn't intuitive? I mean, does anyone really need to consult the chart to know that they're in one of the red boxes - I kind of doubt it. The blue and grey boxes are a different story, but can there really be an argument for pinch hitting for guys (other than your pitcher) in the 3rd or 4th inning. I mean, you'd run through your bench, pretty fast.
Anyway, am just curious to learn if there was more coming - like what to do when you're in a particular colored box, so the practical value can be perceived. In contrast, your earlier, similar piece on when to walk Bonds seemed eminently practical, I think, because the action and its immediate, direct consequences were much more certain.
December 3, 2002 - BeanJTSnow
What are your definitions of 'Very high-leverage', 'High Leverage', etc.?
December 4, 2002 - tangotiger
(www)
(e-mail)
...but the shading doesn't print for me. Just pages of empty grids.
Hmmm... maybe I should put text and color? I'll see what I can do about that.
In some innings, does the blue shading bleed into the -4/+4 run columns (or even further)?
Good question! I was thinking about that, but since I used the same program as for Bonds, I limited to -3/+3. Maybe next time I'll expand to something larger.
...although the column headings would be better if repeated every half-inning, not just every inning.
Thanks for the suggestion! My artistic skills are not what even an average person has, so any formatting improvements suggestions are appreciated. I'll do this next time as well.
But is there anything here that isn't intuitive? I was a bit surprised by how much the leverage changes as soon as you get one guy on base, especially the late innings.
...but can there really be an argument for pinch hitting for guys (other than your pitcher) in the 3rd or 4th inning. I mean, you'd run through your bench, pretty fast. I mentioned at the end that that was not what I was suggesting. Though I would consider this if my batter was Ordonez, and Piazza had the day off.
...like what to do when you're in a particular colored box, so the practical value can be perceived. In contrast, your earlier, similar piece on when to walk Bonds seemed eminently practical, ...
There's really no end to this WE stuff. Eventually, I will be producing charts for the SB break-even points, when to bring in your reliever, should you go for the DP or try the runner at home, should you test the RF's arm, etc, etc. Any suggestions you can offer would be appreciated as well.
What are your definitions of 'Very high-leverage', 'High Leverage', etc.?
It gets a bit dry (series of math equations), but I just picked some arbitrary threshholds to try to distinguish easily the various situations. I could have put +.054 wins and +.013 wins, etc, but who the heck knows what that means?
December 4, 2002 - Matt Rauseo
(www)
(e-mail)
Tango,
Maybe to some readers it may be dry math, but personally I want to know what the thresholds are, I would like to see what exactly is a high leverage situation etc...
Also, I would also like to reiterate, that this would be infinitely more valuable if their was some way of making this printer friendly for those of us using black and white.
Thanks!
December 4, 2002 - Tracy Mohr
(e-mail)
I second Matt's comment above. It may be dry math to some, but what better way to receive feedback, positive or negative, than to post the math (or a link to it)?
December 4, 2002 - Chris Sjoholm
One thing that seemed counter-intuitive to me was that runners on, one out is higher-leverage than runners on, two outs. It seems to me that when sportswriters talk about clutch situations, they more often pull out the "hitting with two outs" stats. But with one out, you get more chances to change the outcome of the game, and are more likely to score. Is that the reason?
December 4, 2002 - tangotiger
(www)
(e-mail)
I'll do my best.
This is what you do: 1 - Determine the WE for every inning/game/base/out for an average team. I've provided a subset of that in the initial link.
2 - Assume that your "great pitcher" or "great hitter" or whatever is going to come in for 1 PA. What is the expected WE following this player's PA? (I used a player whose component stats translates to a .750 win%)
3 - Take the difference between the two. That is the impact in wins of a "typical super-great" player for 1 PA.
The biggest swing, in this example, is about +.07 wins, and that occurs in the bottom of the 9th, home team up by 1, and you have men on 2b and 3b and 1 out. That is, if you bring in say Pedro or RJ or Mo Bonds or Thome or Giambo for ONE SINGLE PA, he will have an effect of .07 wins (assuming these guys are .750 players) over an average player.
How much is +.07 wins? Well, the typical star is +6 wins in 600 PA (+.01 wins). If you bring in Giambi IN THIS PARTICULAR SITUATION 100 times, he'll have as much impact as playing full time.
Now, now, you won't have this situation 100 times, and not having Giambi regularly in the lineup might even mean you might have this situation zero times, who knows. But this is the magnitude of the impact.
So, while Theo Epstein and Bill James are saying that tied games in late innnings are very important (AND THEY ARE!), my research shows that up by 1 for the fielding will have more of an impact to have a great pitcher pitching.
Anyway, the thressholds I used are .01 / .02 / .04. Just made them up to try to get a balance to the chart. Well, I used the .01 because that's what a great player is worth randomly. And .04 cause that would make it 4 PAs in a game. So, given the choice to hit Piazza 4 times randomly, or once in the "very high-leverage" situation, it's a wash. Of course, if that situation doesn't come up, well, you lose on the deal.
Is that enough detail? Too much?
December 4, 2002 - tangotiger
(www)
(e-mail)
Chris, you got it!
If you followed my "Runs Created" series, it shows that the "run environment" (really WIN environment) already exists when the batter/pitcher matchup comes up. That is, the runners on base already have a built-in chance of scoring, given the environment they are playing under.
So, if you then introduce a great player into the mix at that point in time, the entire environment changes. Now, the chances of winning change (sometimes drastically). With 1 out, more damage can be done (not only with the runner on base, but with the batter getting on base). You bring Bonds as a PH with 1 out, not only is the guy on base likely to score, but Bonds will now put himself in a position to extend the inning.
December 4, 2002 - Oliver
Terrific stuff. How often do crucial situations arise in circumstances where strategic options are available (e.g. when the pitcher is at the plate), and what are the costs and benefits of the strategic options? In other words, if used properly, how many games per year is this information on crucial situations worth?
December 4, 2002 - tangotiger
(www)
(e-mail)
Oliver, I really don't know. You'd really have to compare how teams should make their choices optimally against how they really make their choices. And you'd have to break it down by the kinds of choices as well (steals, sacs, taking extra base, throwing to wrong base, bringing in the wrong reliever, batting order, etc, etc, etc). It's gotta be a few wins at least. I don't know, 5? 6?
In the business world, I would perform a cost/benefit analysis. But, the reason I'm doing all this baseball stuff is so that I can get away from doing these boring dry cost/benefit reports! Please don't make baseball like a job for me!!
December 4, 2002 - fracas
This may allow you to print the chart colors (using IE 6):
Tools Internet Options Advanced tab scroll down to Printing Check the box for "Print background colors and images" remember to hit the Apply button
December 4, 2002 - fracas
Oh sure, just when I try to give internet advice, some evil piece of software takes out the carriage returns and makes me look like an idiot. Try this:
Tools
Internet Options
Advanced tab
scroll down to Printing
Check the box for "Print background colors and images"
remember to hit the Apply button
December 5, 2002 - tangotiger
(www)
(e-mail)
Here's a printer-friendlier version
http://groups.yahoo.com/group/tangotiger/files/crucialpa.pdf
Still no text, though.