Some Observations On Phil Hughes

Written by Josh W. | 12 August 2010

Decrease In Strikeouts

From April through May, Phil struck out batters in about 25% of plate appearances. Since that early season stretch of dominance, his K/PA % has dwindled to 18%. This could be due to a multitude of factors, like raw stuff, command, approach, etc.

April-May

screenshot_phil_hughes_march_through_may

June-August

screenshot-1-phil_hughes_june_through_aug

Using these two charts (found using Trip Somers' pitch f/x tool), we can see if there have been any meaningful differences to Phil's raw stuff. FF = fourseam, CU = curveball, FC = cutter, FT = two seam, CH = changeup. Obviously the most important pitches here are Phil's FF, CU, and FC, the three pitches he throws 92% of the time.

His four-seam fastball is nearly idential in both time periods. It was slightly better in April-May, but by a very small margin. The cutter is also very similar in both stretches, being slightly faster since June, and also with a little more vertical movement (relative to a ball thrown without spin). The key here though with the fastball and cutter is that the two are more similar in June-August than April-May. The velocity separation is .6 mph less, and the vertical movement separation .58 inches less. This is important because the qualities of both of these pitches are inextricabley linked; they each make eachother more effective. It is possible that the two pitches have fared worse recently because they are more similar (albeit slightly).

*it is also important to note that it's possible this just a pitch f/x artifact, that the system is having more trouble classifying his pitches recently than before. I'm not sure how likely this is, given that pitch f/x classifications are supposed to get more accurate over time as the system gets more information about a player.

Phil's curve, despite being thrown half a MPH slower in June-August than April-May, lost some of its movement; it lost .41 inches of vertical movement and .34 inches of horizontal movement. This may be also due to the fact that Phil may have tweaked how he throws his curve; that is, how often he throws it as a get-me-over pitch instead of a pitch meant to induce a whiff. It's also possible the pitch was just sharper earlier in the season.

Overall, there are very few differences in Phil's raw stuff in the two stretches. This means that we can likley conclude that the variation in his strikeout rate is due to factors besides raw stuff. One of the largest factors here seems to be approach; both Phil's approach to the league and vice versa.

To examine this it is useful to look at pitch results:

April-May

phils_pitch_resuls_march-may

June-August

phils_pitch_results_may-aug

As you can see, Phil's has significantly altered his repertoire from earlier in the season. He's throwing his cutter much less, and instead is throwing more four-seamers and curves. He also seems less willing to throw strikes with his cutter, probably as a result of him getting beat on it a couple times after that hot start.

Early in the season, hitters only puts Phil's four seam in play 13.4% of the time. Since that point batters have had much less trouble, putting his fastball in play 19% of the time, which is very close to the MLB average of 19.28%. This is because of a change in approach; the league is swinging at Phil's fastball more, and when they do, they are more prepared for the pitch. The league has also swung less at every pitch besides the fastball since April-May. This suggests that opposing batters have decided to sit on his fastball much more often.

Phil's curve has also been whiffed at much less recently. This is likely due a combination of Phil using it earlier in at-bats, and the oppostion having a better approach (which can be seen using Joe Lefkowitz's pitch f/x tool):

April-May

Pitch Type	0-0	0-1	0-2	1-0	1-1	1-2	2-0	2-1	2-2	3-0	3-1	3-2
vs RHB
FC	30.8	31.3	19.6	51.5	36.1	10.5	62.5	63.6	48.5	---	33.3	53.8
FF	54.2	45.3	62.7	36.4	47.2	56.1	25	27.3	24.2	66.7	66.7	15.4
FT	7.5	14.1	7.8	6.1	5.6	10.5	12.5	9.1	12.1	33.3	---	23.1
CU	7.5	9.4	9.8	6.1	11.1	22.8	---	---	15.2	---	---	7.7
vs LHB
FF	69.5	47.1	58.1	68.3	40	45.2	75	65.4	51.9	100	60	38.7
CU	12.7	14.7	23.3	7.3	18	20.5	6.3	---	9.3	---	---	16.1
FC	12.7	26.5	2.3	22	34	23.3	18.8	34.6	29.6	---	40	32.3
CH	0.8	4.4	4.7	2.4	4	2.7	---	---	3.7	---	---	3.2
FT	4.2	7.4	11.6	---	4	8.2	---	---	5.6	---	---	9.7
PO	---	---	---	---	---	---	---	---	---	---	---	---

June-August

Pitch Type	0-0	0-1	0-2	1-0	1-1	1-2	2-0	2-1	2-2	3-0	3-1	3-2
vs RHB
FF	56.2	64.9	74.2	39.3	51.5	64.4	83.3	48.6	71.2	80	66.7	65.4
FC	23.3	16.9	---	41.1	23.5	9.6	11.1	42.9	15.2	---	22.2	23.1
CU	17.1	14.3	19.4	12.5	22.1	23.3	---	8.6	7.6	---	11.1	7.7
PO	---	---	---	---	---	---	---	---	---	---	---	---
FT	3.4	3.9	6.5	5.4	2.9	2.7	---	---	6.1	---	---	3.8
IN	---	---	---	1.8	---	---	5.6	---	---	20	---	---
vs LHB
FF	62.8	57.5	71.1	63.8	44.1	50.9	73.3	60.5	49.1	60	93.3	63.6
FT	5.8	2.5	5.3	6.9	10.3	7	6.7	7.9	3.8	20	---	---
CU	22.4	12.5	18.4	17.2	29.4	28.1	---	10.5	28.3	---	---	24.2
FC	6.4	18.8	5.3	6.9	5.9	10.5	13.3	15.8	15.1	20	6.7	9.1
CH	2.6	8.8	---	3.4	8.8	3.5	6.7	5.3	3.8	---	---	3
??	---	---	---	1.7	1.5	---	---	---	---	---	---	---

It seems that Phil's drop in strikeouts can be primarily explained by a change in approach, and not raw stuff; he's thrownig his cutter much less, and replacing it with pitches that batters have less trouble making contact against. I'd also wager that because Phil is throwing his fastball more, and that because hitters are putting it in play more often, that Phil is getting into less deep counts where he has an oppurtunity to strikeout the hitter, but that's just conjecture at this point.

Pitch Selection Verse Lefties and Righties

Pitch Type	Count	Selection%	Take%	Swing%	Swing-Miss%	HR%	GB%	LD%	FB%
vs RHB
FC	267	26	37.1	62.5	16.2	0.7	35.8	28.4	35.8
FF	558	54.4	43.4	56.5	19.7	0.2	25.8	21.3	52.8
FT	64	6.2	53.1	46.9	16.7	0	57.1	0	42.9
CU	137	13.4	75.9	23.4	9.4	0	50	37.5	12.5
vs LHB
FF	658	57.4	50.3	49.2	15.1	1.4	28.4	20.2	51.4
CU	199	17.4	68.8	30.7	32.8	0.5	61.9	19	19
FC	180	15.7	39.4	59.4	21.5	1.7	30.8	17.9	51.3
CH	42	3.7	69	31	30.8	0	50	0	50
FT	65	5.7	49.2	50.8	12.1	0	50	0	50
??	2	0.2	50	50	0	0	0	0	0
Pitch Type	Count	Selection%	Take%	Swing%	Swing-Miss%	HR%	GB%	LD%	FB%
TOTAL	2172	100	49.7	49.9	18.2	0.7	34.1	21.6	44.3

As you can see, Phil throws fastballs a whopping 86.6% of the time to right handed batters. Against lefties, this number drops to 78.9. For Phil to optimize his pitch selection, he should probably throw his curve more to lefties, and less to righties. The reason why is in bold. His curve is simply ineffective to righties (only 2.2% of total curves against righties are whiffed at); righties convert the pitch into a linedrive 37.5% of the time! Against lefties, however, his curve is much better. As you can tell using the graph, lefties have a contact rate of 67.2% against the curve, better than the major league average of 73.9% (though for righties and lefties).

When righties aren't taking Phil's curve, they are usually putting it in play, and hitting it hard. Against righties, using the linear weights found here, we can find out that the average run value of a BIP hit off the curve is .705, while it's .65 for lefties. From this we can conclude that lefties both have trouble making contact against Phil's curve and when they do make contact they don't hit it as hard as righties.

Is this platoon split unusual? Turns out, not at all. Thanks to Max Marchi, we know that "roundhouse curves", or curves that avg. 75. mph with half of foot of vertical and horizontal movement, having a huge platoon split. Phil has a curve very similar to the type Marchi profiles. Interestingly, curves thrown harder than Phil's (around 81 mph) showcase a much smaller platoon split.

This seems to suggest that Phil should throw his curve less to righties this year, and then in the offseason, work on throwing it harder. Then next season he would have a decent curve against righties.